CN114328600A - Method, device, equipment and storage medium for determining standard data element - Google Patents

Abstract

The present application provides a method, apparatus, device and storage medium for determining a standard data element. The method includes: obtaining at least one characteristic information of a target data element; and determining from a standard data element library based on each characteristic information of the target data element A first data element set matching the target data element is obtained, and a second data element set corresponding to the target data element is determined from the historical benchmarking record library; according to the standardization times of the second standard data element in the second data element set, determine The recommendation score of the second standard data element; combining the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set, from the first data element set and the second standard data element At least one first candidate standard data element for standardizing the target data element is determined in the two data element sets. The solution of the present application can improve the accuracy of standard data element matching.

Description

Method, apparatus, device and storage medium for determining standard data elements

technical field

The present application relates to the technical field of data processing, and more particularly, to a method, apparatus, device and storage medium for determining standard data elements.

Background technique

With the continuous development of informatization and digitization, enterprises can conduct data regulation based on the data center.

In the data normalization stage, data element standardization needs to be carried out. Standardization of data elements refers to converting data elements into standard data elements that conform to industry standards or national standards and other standard specifications. Based on this, in the standardization process of the data elements, it is necessary to determine the standard data elements that match the data elements. At present, it is difficult to accurately match standard data elements suitable for the data elements from the standard data element library, so that the matching accuracy of the standard data elements is low.

SUMMARY OF THE INVENTION

In view of the above problems, the present application provides a method, apparatus, device and storage medium for determining standard data elements, so as to improve the accuracy of standard data element matching. The specific plans are as follows:

In a first aspect of the present application, a method for determining standard data elements is provided, comprising:

Obtain at least one feature information of the target data element to be standardized;

Based on each feature information of the target data element, a first data element set matching the target data element is determined from the standard data element library, where the first data element set includes: Each first standard data element matched by the target data element, and the degree of matching between the first standard data element and the target data element;

Based on the characteristic information of the target data element, the second data element set corresponding to the target data element is determined from the historical benchmarking record library, and the historical benchmarking record library stores: characteristic information of the non-standard data element, at least one standard data element after the non-standard data element has been standardized in the history, and the non-standard data element has been standardized as the normalization times of each standard data element in the history; the The second data element set includes: each second standard data element that has been standardized in the history of the target data element and the normalization times corresponding to the second standard data element;

The recommendation score of the second standard data element is determined according to the normalization times of the second standard data element in the second data element set, wherein, the higher the normalization times of the second standard data element, the higher the second standard data element. The higher the recommendation score of the data element is, the recommendation score of the second standard data element is used to represent the suitability of the second standard data element as the standard data element of the target data element;

Combining the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set, from the first data element set and the second data element set At least one first candidate standard data element for normalizing the target data element is determined.

In a possible implementation manner, the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set are combined, from the first data element set. Determining at least one first candidate standard data element for standardizing the target data element in a data element set and a second data element set includes:

Combining the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set, it is determined that the first data element set and the second data element set are in the The first comprehensive score of each standard data element;

generating a first candidate normalization list for normalizing the target data element, where the first candidate normalization list includes: a first comprehensive score determined from the first data element set and the second data element set is higher of at least one first candidate standard data element.

In another possible implementation manner, the first candidate standardization list further includes: a first comprehensive score of the first candidate standard data element;

The method also includes:

Based on at least one feature information of the target data element, construct a feature word segment set of the target data element, and the feature word set includes at least one feature word segment obtained from the at least one feature information segment;

Based on the feature word segmentation set, determine a third data element set in the standard data element library that is similar to the target data element, where the third data element set includes: the similarity between the feature information set and the feature word segmentation set Higher at least one third standard data element, and the first similarity between the feature information set of the third standard data element and the feature word segmentation set; the feature information set of the standard data element includes the each characteristic information;

Based on the first similarity corresponding to each third standard data element and the first comprehensive score of each first candidate standard data element in the first candidate standardization list, from the third data element set and the first candidate standardization In the list, at least one target standard data element for normalizing the target data element is determined.

In a second aspect of the present application, a device for determining standard data elements is provided, comprising:

an information obtaining unit for obtaining at least one feature information of the target data element to be standardized;

a first set determining unit, configured to determine a first set of data elements matching the target data element from a standard data element library based on each feature information of the target data element, the first set of data elements comprising: Each first standard data element in the standard data element library that matches the target data element, and the degree of matching between the first standard data element and the target data element;

The second set determining unit is configured to determine, based on each feature information of the target data element, a second data element set corresponding to the target data element from the historical benchmarking record library, and the historical benchmarking record library stores There are: characteristic information of the non-standard data elements that have been standardized in the history, at least one standard data element that has been standardized in the history of the non-standard data elements, and, the non-standard data elements have been standardized into each standard in the history. Standardization times of data elements; the second data element set includes: each second standard data element that has been standardized in the history of the target data element and the standardization times corresponding to the second standard data element;

A recommendation scoring unit, configured to determine the recommendation score of the second standard data element according to the normalization times of the second standard data element in the second data element set, wherein the more the normalization times of the second standard data element , the higher the recommendation score of the second standard data element is, the recommendation score of the second standard data element is used to represent the suitability of the second standard data element as a standard data element of the target data element;

The first data element determining unit is configured to combine the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set, from the first data element set. At least one first candidate standard data element for normalizing the target data element is determined in the meta set and the second data element set.

In a third aspect of the present application, a computer device is provided, comprising: a memory and a processor;

the memory for storing programs;

The processor is configured to execute the program to implement each step of the method for determining a standard data element in any embodiment of the present application.

In a fourth aspect of the present application, a storage medium is provided on which a computer program is stored, characterized in that, when the computer program is executed by a processor, the method for determining a standard data element in any one of the embodiments of the present application is implemented. of the various steps.

With the above technical solution, the present application will not only search for the first standard data element matching the target data element from the standard data element library, but also determine that the target data element has been standardized historically based on each feature information of the target data element. Each subsequent second standard data element and the target data element are normalized to the normalization times of each second standard data element. Since the normalization times of the second standard data element reflects the number of times the target data element has been normalized to the second standard data element in history, the second standard data element determined based on the normalization times of each second standard data element The recommendation score of can reflect the suitability of the second standard data element for normalizing the target data element. On this basis, the present application combines the matching degree of each first standard data element with the target data element and the recommendation score of each second standard data element to determine the standardization of the target data element, taking into account the characteristic information of the target data element The matching situation with the standard data element, and considering some behavior information that has been standardized in the history of the target data element, the standard data element suitable for standardizing the target data element can be more accurately determined, and the recommendation can be more accurate. The standard data element used to standardize this data element.

Description of drawings

Various other advantages and benefits will become apparent to those of ordinary skill in the art upon reading the following detailed description of the preferred embodiments. The drawings are for purposes of illustrating preferred embodiments only and are not to be considered limiting of the application. Also, the same components are denoted by the same reference numerals throughout the drawings. In the attached image:

1 is a schematic flowchart of a method for determining a standard data element provided by an embodiment of the present application;

2 is a schematic diagram of a storage structure for storing standard data elements to which non-standard data elements are historically standardized in a historical benchmarking record library provided by an embodiment of the present application;

3 is a schematic flowchart of a method for determining a standard data element provided by an embodiment of the present application;

4 is a schematic diagram of an implementation process of constructing a feature word segmentation set based on at least one feature information of a target data element provided by an embodiment of the present application;

FIG. 5 is a schematic flowchart of an implementation of determining an overall score of a candidate standard data element according to an embodiment of the present application;

6 is a schematic diagram of an implementation principle framework of a method for determining a standard data element provided by an embodiment of the present application;

7 is a schematic structural diagram of a composition of an apparatus for determining a standard data element provided by an embodiment of the present application;

FIG. 8 is a schematic diagram of a composition structure of a computer device to which the solution according to the embodiment of the present application is applied.

Detailed ways

Different industries or different fields of the same industry may name the same data element differently. When it is necessary to use the data center and other technologies to manage the same data, the data elements need to be unified and standardized, which requires at least one standardized data element that can be used to standardize the data element, so that the user can choose a standard data element as the The normalized data element of the data element.

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, but not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

The inventors of the present application have found through research on the matching process of standard data elements: for the data element to be standardized, the characteristic information of the data element can be used to match the standard data element from the standard data element library. However, this method can only match a standard data element that is strongly related to the characteristic information of the data element. If the characteristic information of the data element to be standardized is a common word, it is difficult to accurately match a suitable standard data element.

In order to match the standard data elements more accurately and reduce the situation that the standard data elements cannot be matched, the present application also matches the standard data elements from the standard data element library based on the characteristic information of the target data elements to be matched. Combined with the feature information of the target data element, each standard data element into which the target data element has been standardized in the history is detected from the pre-built historical benchmarking record library. Combined with the detection results of the standard data element library and the historical benchmarking record library, the standard data element used to standardize the target data element is comprehensively determined, so that the standard data element can be comprehensively determined in combination with the matching results of multiple dimensions, which can also be updated. In order to accurately determine the standard data element suitable for the standardization target data element.

The method for determining standard data elements provided by the present application will be introduced below with reference to the flowchart.

As shown in FIG. 1 , which shows a schematic flowchart of a method for determining a standard data element provided by an embodiment of the present application, the method in this embodiment may include:

S101, at least one characteristic information of the target data element to be standardized is obtained.

A data element refers to a unit of data maintained by different industries. For example, the data elements can be names, ID numbers, postal codes, and so on.

It can be understood that, in order to facilitate the distinction, this application refers to the data elements that need to be standardized as target data elements.

The feature information of the target data element may include the name of the data element, such as one or both of the English code and the Chinese name of the data element. The feature information of the data element may also be description information used to characterize the features of the data element, such as one or more of the data type of the data element, the description information of the data element, and the industry to which the data element belongs.

In this application, at least one feature information of the target data element includes at least the name of the data element, such as the English code of the data element, and may also include other features of the data element, which is not limited.

S102, based on each feature information of the target data element, determine a first data element set matching the target data element from a standard data element library.

The first data element set includes: each first standard data element matched with the target data element in the standard data element library, and the matching degree between the first standard data element and the target data element.

The standard data element library may store respective characteristic information of a plurality of standard data elements. The characteristic information of the standard data element may also include the English code and Chinese name of the standard data element and other identifiers, and may also include one or more of the description of the standard data element, the industry to which it belongs, the representation format, and the category.

For example, the standard data element library may record the characteristic information of each standard data element in the form of Table 1 below.

Table 1

In Table 1, each row corresponds to the feature information of a standard data element.

Among them, each standard data element has an English code and a Chinese name. For example, the first row in Table 1 records a standard data element whose English code is "GMSFHM", and the Chinese name of the standard data element is "citizen ID number". At the same time, the standard data element library also has descriptions, synonyms, data types of standard data elements, object categories represented by standard data elements, representation formats and code length labels of standard data elements.

The object categories may include: people, roads, time, documents, vehicles, etc., which are not limited.

The representation format is the length of digits used to represent standard data elements, for example, the representation format of a citizen ID number is 18 digits.

Since the representation format is too refined, it is easy to cause confusion when matching data elements to be standardized. This application also proposes to use a code length label to mark the length of standard data elements. In this application, the code length labels can be divided into two categories, one is a fixed code length label, and the other is a dynamic code length label. The fixed code length label shows that the standard data element has a specific code length (such as ID card, license plate number, mobile phone number, etc.), type value length, time field length, etc. The dynamic code length label is expressed as the code length floating in a certain interval (such as person's name, address, description, etc.).

Of course, the dynamic code-length sticky note can be further divided into narrow code and wide code, and the division method can divide the confidence value based on the length distribution of the data element samples. The code is the length of the code length note> 10.

It can be understood that, for each feature information of the target data element, the standard data elements can be matched from the standard database metabase respectively, and finally the standard data elements to which at least one feature information of the target data element is respectively matched can be obtained. This application does not limit the specific implementation of matching standard data elements from standard database elements.

For the convenience of distinction, the standard data element matched from the standard data element library is referred to as the first standard data element in this application.

Wherein, the first standard data element matched with the target data element may be: a standard data element whose matching degree between the characteristic information and one or more characteristic information of the target data element exceeds the set threshold; or, belongs to the characteristic information and the target data element The standard data element is set for the higher matching degree of the characteristic information of the data element.

The degree of matching between the first standard data element and the target data element is the similarity between the characteristic information of the first standard data element and the characteristic information of the target data element.

S103, based on each feature information of the target data element, determine a second data element set corresponding to the target data element from the historical benchmarking record library.

The historical benchmarking record library stores: characteristic information of the non-standard data elements that have been standardized in the history, at least one standard data element that has been standardized in the history of the non-standard data elements, and, respectively, the non-standard data elements that have been historically standardized Normalization is the number of normalization times for each standard data element.

The non-standardized data elements are historically obtained data elements that need to be standardized.

The standard data element to which the non-standard data element has been standardized in the past may be: the standard data element to which the non-standard data element has been standardized by the user in the history. Correspondingly, the historical benchmarking record library can be constructed based on the behavior records of standard data elements selected for the data elements to be standardized by different users in history.

It can be understood that, due to different business scenarios, for the same data element, the standard data element selected by the user for the data element may also be different. In addition, due to other reasons such as misoperation, the user may also select different standardized data elements for the same data element in different standardization processes. Based on this, each non-standard data element in the historical benchmarking record library may correspond to multiple standard data elements, and the normalization times corresponding to each standard data element are different.

For example, at least one standard data element corresponding to the English code (or Chinese name) of each non-standard data element and the standardization times of each standard data element may be recorded in the historical benchmarking record library. In the historical benchmarking record library, each standard data element corresponding to the non-standard data element is a standard data element that the non-standard data element has been selected and standardized into in the history. use

For example, the historical benchmarking record library can record each non-standard data element and its historically standardized information in the form of FIG. 2 .

As can be seen from Figure 2, a non-standard data element can correspond to one or more standard data elements that have been standardized in history, and the number after each standard data element is the standard data element that was selected to be standardized to the standard in history. The number of data elements. As shown in FIG. 2 , the number of times the non-standard data element is normalized to the standard data element 1 in the history is 2 times, and the number of times that the standard data element 2 is selected to be normalized in the history is 5 times.

It can be understood that the present application can combine the types of feature information of non-standard data elements recorded in the historical benchmarking record library, and combine the feature information of the corresponding types of target data elements with the characteristics of each non-standard data element in the historical benchmarking record library. information is compared. For example, if at least the English code of the non-standard data element is recorded in the historical benchmarking record database, then at least the English encoding of the target data element can be matched with the English encoding of each non-standard data element in the historical benchmarking record database to match the English code of each non-standard data element. The second set of data elements.

In this application, the second data element set includes: each second standard data element after the target data element has been standardized in history and the normalization times corresponding to the second standard data element.

For the convenience of distinction, the standard data element matched from the historical benchmarking record library is called the second standard data element, but it is understandable that the standard data element matched from the historical benchmarking record library may be different from the standard data element from the standard data element. The matched standard data elements in the metabase are repeated, which may cause a second standard data element and a first standard data element to be the same standard data element.

S104, according to the normalization times of the second standard data element in the second data element set, determine the recommendation score of the second standard data element.

Wherein, the more normalization times of the second standard data element, the higher the recommendation score of the second standard data element.

It can be understood that the more times the second standard data element is standardized, the more times the second standard data element is selected by the user as the standard data element, which means that the second standard data element is more suitable for standardized correspondence. of unnormalized data elements. Based on this, the more times the second standard data element is standardized, the more suitable the second standard data element is as the standardized data element of the standardization target data element, and therefore, the higher the recommendation score is. It can also be known that the recommendation score of the second standard data element can be used to represent the suitability of the second standard data element as the standard data element of the target data element.

It can be understood that, for a second standard data element corresponding to a non-standard data element, the time when the second standard data element was last selected as a standardized data element is closer to the time at the current moment, the second standard data element. The possibility that the element is re-selected as the normalized data element of the non-standard data element is also higher, that is to say, the second standard data element is more suitable for normalizing the non-standard data element (or the target data element). Based on this, in the present application, the historical benchmarking record library also stores: the standard time when each non-standard data element was last normalized to each standard data element in the history.

For the convenience of distinction, the time when the non-standardized data element is last normalized to a certain standard data element in the history is referred to as the last normalization time of the standard data element as the standard data element of the non-standardized data.

Correspondingly, the second data element set may further include the latest standardized time corresponding to the second standard data element. On this basis, the present application may further determine the recommendation score of the second standard data element according to the standardization times of the second standard data element in the second data element set and the latest standardization time. Wherein, the more the normalization times of the second standard data element and the shorter the duration from the last normalization time to the current time, the higher the recommendation score of the second standard data element.

There are various possibilities for the specific manner of determining the recommendation score of the second standard data element, which is not limited.

For ease of understanding, a possible situation for calculating the recommendation score of the second standard data element is described below as an example:

If the determined number of the second standard data element is only one, the recommendation score of the second standard data element may be set to 100;

If there are multiple second standard data elements determined:

For the second standard data element with the largest number of normalization times among the plurality of second standard data elements, the second standard data element may be set to 100 points;

For each second standard data element whose normalization times are not the most among the plurality of standard data elements, the following formula 1 can be used to calculate the recommendation score Y of the second standard data element:

Y=119.91X ^-0.286 *e ^{(-λ*max(0,td))} (Formula 1);

Wherein, X is the difference between the normalization times of the second standard data element and the maximum normalization times, and the maximum normalization times is the maximum value among the normalization times of multiple second standard data elements matched by the target data element; λ is the set Based on the rate of time fraction decay, it can be set according to actual needs; t is the difference in days between the last normalized time corresponding to the second standard data element and the current day, d is the set number of days in the time interval that does not require decay, For example, if it is assumed that 7 days can be ignored due to the influence of time on the recommendation score, the value of d is 7.

Of course, here is an example of a method for calculating the recommendation score, and the calculation of the recommendation score by other methods is also applicable to this application, and there is no restriction on this.

S105: Determine the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set, from the first data element set and the second data element set. at least one first candidate standard data element in the standardization target data element.

Wherein, for the convenience of distinction, the standard data element selected from the first data element set and the second data element set is referred to as the first candidate standard data element.

It can be understood that the matching degree and recommendation score of the standard data element essentially reflect the suitability of the standard data element for normalizing the target data element. Moreover, in many cases, the standard data elements in the first data element set and the second data element set may overlap. Based on this, considering whether each standard data element in the first data element set and the second data element set exists in the two data element sets at the same time, and at least one of the matching degree and recommendation score of each standard data element, it is possible to The standard data elements that are more suitable or more likely to be selected by the user for standardization in the first data element set and the second data element set are comprehensively determined as candidate standard data elements.

There are many possibilities for the specific implementation manner of determining the first candidate standard data element, which is not limited.

For example, in a possible implementation manner, the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set may be combined to determine the first standard data element. The first comprehensive score of each standard data element in the data element set and the second data element set. Then, combined with the first comprehensive score of each standard data element, a first candidate normalization list for normalizing the target data element is generated.

Wherein, the first candidate standardization list includes: at least one first candidate standard data element with a higher first comprehensive score determined from the first data element set and the second data element set.

Of course, as an optional manner, the first candidate standardization list may further include: a first comprehensive score of each first candidate standard data element. Based on this, the at least one first candidate standard data element can be sorted and output in combination with the first comprehensive score of each first candidate standard data element, so that the user can select the required standard data element.

Wherein, for any standard data element in the first data element set and the second data element set, if the standard data element exists in both the first data element set and the second data element set, then the first data element of the standard data element The comprehensive score is related to the matching degree and recommendation score of the standard data element; if the standard data element only belongs to the first data element set or the second data element set, the first comprehensive score of the standard data element is only related to its matching degree or recommendation score .

Certainly, there may be many specific ways of determining the first comprehensive score. The following is an example of a possible implementation:

First, the matching degree and recommendation score can be unified into the same numerical range. For example, take the matching degree and the recommendation score as percentage values as an example. For any first standard data element, if the matching degree is a probability value belonging to 0-1, the matching degree can be standardized first.

For example, the matching degree of the first standard data element can be standardized by the following formula 2, and the matching degree score' after the matching degree score of a standard data element is normalized:

Wherein, max(score) is the maximum matching degree of all the first standard data elements in the first data element set; min(score) is the minimum matching degree of all the first standard data elements in the first data element set. Of course, if the matching degree corresponding to the first standard data element is determined, if the standardized matching degree is taken as the matching degree of the first standard data element, it is not necessary to perform the processing of standardizing the matching degree.

Secondly, for any standard data element in the first data element set and the second data element set, for the matching degree (standardized matching degree) of the standard data element and any score socre(i) in the recommendation score, we can Calculate the score socre(i) and score Item_socre(i) after stimulation by the following formula 3:

Item_socre(i)=socre(i)+(max_socre-socre(i))*hitBoost (formula 3);

Among them, max_socre is the set top score, and the top score can be 100 when both the standardized matching degree and the recommendation score are based on a percentage system. The hitBoost is a set incentive weight, wherein the same incentive weight can be used for the matching degree and the recommendation score, or the incentive weights of the matching degree and the recommendation score can be set separately as required.

Finally, for any standard data element in the first data element set and the second data element set, the average value of all post-stimulation scores corresponding to the standard data element can be obtained, and the obtained score value can be used as the standard data element the first overall score.

For example, if the standard data element belongs to both the first data element set and the second data element set, then the standard data element has the corresponding matching degree and recommendation score at the same time, then the post-stimulation score and recommendation score of the matching degree of the standard data element can be calculated The average value of the post-incentive scores was obtained to obtain the first comprehensive score. If the standard data element only belongs to the first data element set or the second data element set, then the first comprehensive score of the standard data element is the post-stimulation score of the matching degree or the post-stimulation score of the recommendation score.

It is understandable that, considering that matching the standard data element directly from the standard data element library based on the characteristic information of the target data element is a relatively accurate data element matching method, in practical applications, the user can also set the standard data element according to their needs. Determines whether to enable the option to directly recommend individual standard objects. If the user chooses to enable the option of direct recommendation of a single data element, and a unique and identical standard data element is matched from the standard data element library and the historical benchmarking record library, then the standard data element can be directly determined as the first candidate standard data element, then the operation of calculating the first comprehensive score is no longer required.

It can be seen from the above that the present application not only searches the standard data element library for the first standard data element matching the target data element, but also determines the historically standardized value of the target data element based on the characteristic information of the target data element. Each second standard data element and the target data element are normalized to the normalization count of each second standard data element. Since the normalization times of the second standard data element reflects the number of times the target data element has been normalized to the second standard data element in history, the second standard data element determined based on the normalization times of each second standard data element The recommendation score of can reflect the suitability of the second standard data element for normalizing the target data element. On this basis, the present application combines the matching degree of each first standard data element with the target data element and the recommendation score of each second standard data element to determine the standardization of the target data element, taking into account the characteristic information of the target data element The matching situation with the standard data element, and considering some behavior information that has been standardized in the history of the target data element, the standard data element suitable for standardizing the target data element can be more accurately determined, and the recommendation can be more accurate. The standard data element used to standardize this data element.

In addition, since the target data element is a non-standardized data element, the format or content of the feature information of the target data element is not standardized, which may lead to the fact that the feature information based on the target data element cannot be accurately matched or even from the standard data element library. is unable to match standard data elements. However, since the feature information of non-standard data elements is recorded in the historical benchmarking record library, based on the feature information of the target data element, it is easier to accurately locate the matching non-standard data elements from the historical benchmarking record library, and then Obtain the information of the standard data elements to which the non-standard data elements matched by the target data elements are historically normalized to reduce the problem that the standard data elements for standardization cannot be determined because the characteristic information of the target data elements cannot be covered by the standard data element library. Happening.

It can be understood that, in the above embodiments of the present application, the first candidate standard data element is obtained by directly matching the standard database element library and the historical benchmarking record library by using the feature information of the target data element, that is, based on the target data. The feature information of the meta is accurately matched.

However, in practical applications, due to various reasons such as the irregularity of the feature information of the target data element, the feature information based on the target data element may not be able to match the standard data element from the standard data element library and the historical benchmarking record library. . Moreover, if the target data element has not been standardized in history, the historically adopted standard data element may also not be matched from the historical benchmarking record library.

Based on this, in order to improve the probability of matching from the standard data element library to the standard data element based on the feature information of the target data element, the present application also provides a method for fuzzy matching to the standard data element library based on the feature information of the target data element. On this basis, the present application can comprehensively determine a candidate standard data element for standardizing the target data element in combination with the standard data elements matched in the previous precise matching and fuzzy matching.

Further, the present application can also match the standard data elements corresponding to different non-standard data elements in the historical benchmarking record library based on the feature information of the target data elements, so as to obtain the matched standard data elements, and then combine precise matching and fuzzy matching. For each standard data element obtained, a standard data element for standardizing the target data element is finally determined.

The following describes the process of combining exact matching and fuzzy matching to determine standard data elements with reference to the flowchart.

As shown in FIG. 3 , which shows another schematic flowchart of a method for determining a standard data element in the present application, the method in this embodiment may include:

S301, at least one characteristic information of the target data element to be standardized is obtained.

S302, based on each feature information of the target data element, determine a first data element set matching the target data element from the standard data element library.

The first data element set includes: each first standard data element matched with the target data element in the standard data element library, and the matching degree between the first standard data element and the target data element.

S303, based on each feature information of the target data element, determine a second data element set corresponding to the target data element from the historical benchmarking record library.

The second data element set includes: each second standard data element that has been standardized in the history of the target data element and the normalization times corresponding to the second standard data element.

S304: Determine the recommendation score of the second standard data element according to the normalization times of the second standard data element in the second data element set.

Wherein, the more normalization times of the second standard data element, the higher the recommendation score of the second standard data element.

S305, combining the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set, determine the first data element set and the second data element set. The first overall score for each standard data element.

S306, generating a first candidate standardization list for standardizing the target data element in combination with the first comprehensive score of each standard data element.

The first candidate standardization list includes: at least one first candidate standard data element with a higher first comprehensive score determined from the first data element set and the second data element set, and each first candidate standard The first composite score for the element.

The above steps S302 to S306 are actually the process of accurately matching the standard data element database and the historical benchmarking database based on the feature information of the target data element to obtain candidate standard data elements.

For the above steps S301 to S306, reference may be made to the relevant introductions in the previous embodiments, which will not be repeated here.

S307, based on at least one feature information of the target data element, construct a feature word segmentation set of the target data element.

Wherein, the feature word set includes at least one feature word segmented by at least one feature information word segment.

The purpose of performing word segmentation on at least one feature information is to reduce the situation that a single feature information as a whole cannot match standard data elements with similar features.

For example, in a possible implementation manner, at least one feature information of the target data element may be combined into a text; then, the text is segmented to obtain a feature segment consisting of at least one feature segment obtained from the text segment. gather.

Further, considering that the format of the feature information of the target data element may not be standardized, it is not conducive to hitting the feature information of the standard data element from the standard data element library. Before combining into text, it is also necessary to standardize the feature information of the target data element, and then perform word segmentation and build a feature word segmentation set.

Figure 4 is a schematic diagram of an implementation process of constructing a feature segmentation set based on at least one feature information of a target data element in this application, and the process includes:

S41 , for each feature information of the target data element, based on a plurality of standard features stored in a standard feature library, determine a standard feature matching the feature information, and standardize the feature information based on the standard feature.

The standard feature library may be a corpus constructed based on the collected standard feature names of different scenarios and industries. Therefore, the standard feature library stores standard names of different feature information.

In an example, the multiple standard features in the standard feature library can be further divided into three types, namely: trunk type, branch type, and leaf type.

Wherein, for the standard feature of the trunk type Trunk, if the feature information of the target data element matches the standard feature, the feature information will be completely replaced with the standard feature information. For example, in the definition of "name", the Chinese name fields of data elements in different business scenarios include: "name", "nickname", "flower name", "nickname", etc., but the semantics of the description are the same, then These feature information are replaced with a unique standard semantic feature "name".

For branch-type Branch, the standard feature is a feature with a modified noun. When the standard feature matching is performed based on the feature information of the target data element, it can be detected whether the feature information of the target data element contains a part that matches the standard feature of this type. Feature information, if it exists, replace this part of the feature information with standard features. For example, for the feature information of "the Chinese name of the administrative division of the place of origin", the standard features "implementation code" and "name" need to be matched to the standard features through inclusion rules, then only the "administrative division" needs to be replaced with "city and county code", "Chinese "Name" is replaced with "Name", and finally standardized to "city and county code name of the place of origin".

For the standard features of leaf-type Leaf, after the feature information of the target data element matches the standard features of this type, the suffix words in the feature information can be expanded, for example, "xxxid" is adapted to the ending "id" and expanded to "xxxid primary key" ".

As shown in Table 2 below, Table 2 is information related to various types of standard features in the standard feature library.

As shown in Table 2, the second column is the name of each standard feature, and the first column is the description of the non-standard feature information corresponding to the standard feature. If the feature information of the target data element matches the non-standard feature information of the standard feature, you can Based on the standard feature, the feature information of the target data element is standardized.

Among them, the part-of-speech tag represents the category of each standard feature.

In this application, for each feature information of the target data element, standard feature matching may be performed in the order of Trunk type, Branch type, and Leaf type. Specifically, it is possible to first detect whether there is a standard feature matching the feature information in the standard features of the trunk-type Trunk, and if so, standardize the feature based on the matched standard feature; if the standard features of the trunk-type Trunk do not exist If there is a standard feature that matches the feature information, continue to monitor whether there is a standard feature that matches the feature information in the standard features in the Branch type. If it still does not exist, then check whether there is a standard feature in the Leaf type that matches the feature information. matching standard features.

It can be understood that at least one standardized feature information corresponding to the target data element can be finally obtained through step S41.

S42, combine at least one feature information of the target data element into text.

S43, perform word segmentation on the text to obtain a characteristic word segmentation set consisting of at least one characteristic word segmentation obtained from the word segmentation of the text.

For example, assuming that the English code of the target data element "payment amount" is "payNum", and the Chinese name is "payment amount", the characteristic information of the target data element is standardized and cleaned, and finally a text is formed as "payment amount and amount amount narrow code payNum" ”, the text is segmented, and multiple feature segments can be obtained, and these multiple feature segments can form a feature segment set. For example, the feature segment set can be {payment amount, amount, amount, narrow code paynum}, the feature segment set It can be regarded as a vector composed of multiple feature word segmentations.

Wherein, there may be many ways to segment the text. For example, the present application may use the IK word segmentation method to perform word segmentation, etc., which is not limited.

S308, based on the feature word segmentation set, determine a third data element set in the standard data element library that is similar to the target data element.

Wherein, the third data element set includes: at least one third standard data element with a high similarity between the feature information set and the feature word segmentation set, and the first data element between the feature information set of the third standard data element and the feature word segmentation set similarity.

Wherein, the characteristic information set of the standard data element includes each characteristic information of the standard data element.

The calculation of the similarity between the feature word segmentation set and the feature information set of the standard data element can be calculated by using a commonly used algorithm for calculating similarity. For example, a vector space model (VSM) model or a BIM model can be used for calculation. Similarity, etc., there is no limit to this.

It can be understood that, since the feature word segmentation set includes each word segment after at least one feature information combination and word segmentation, the similarity between the feature word segmentation set composed of these word segmentations and the set composed of each feature information of the standard data element is selected to select. The standard data element is not an exact match of the standard data element, but it can reduce the situation that the standard data element cannot be matched because the feature information of the target data element is directly matched with the characteristic information of the standard data element.

It can be understood that, considering that there are many standard data elements and their characteristic information in the standard data element library, which is not conducive to relatively quickly retrieving each third standard data element, the application can also pre-build the standard data element library corresponding standard data element index library.

The standard data element index library can store data element index files of different standard data elements. The feature information of the standard data element recorded in the standard data element library is stored in the data element index file of the standard data element.

In one example, the present application may use the Lucene engine to index each standard data element item of the standard data element library. Specifically, it can be as follows:

First, for each standard data element, the English code of the standard data element is used as the file name of the data element index file of the standard data element, and the "English code", "Chinese name", "synonym" of the standard data element And feature information such as "description" as the content item of the data element index file.

Secondly, create a corresponding index for the data element index file of each standard data element through Lucene, wherein the file name and file content of the data element index file are written into the index library as different fields in the index.

Among them, the standard data element index library can use the IK word segmentation engine to perform index query.

On the basis of the constructed standard data element index library, this embodiment can directly determine the data element index file from the standard data element index library based on the data element index files corresponding to different standard data elements stored in the standard data element index library A third set of data elements similar to the feature word segmentation set.

Correspondingly, the third data element set includes: at least one third standard data element with a high similarity between the data element index file and the characteristic word segmentation set, and the data element index file of the third standard data element and the characteristic word segmentation set. the first similarity.

It can be understood that, because each feature of each standard data element is used as the content item of the data element index file in the standard data element index library, the index query of each data element index file can be directly based on the feature word segmentation set, which is conducive to improving the matching. speed, so that the third data element set can be determined relatively quickly.

It should be noted that, in this embodiment, fuzzy matching is performed on the standard data element library and the historical benchmarking record library simultaneously based on the characteristic word segmentation set of the target data element as an example for description. However, it can be understood that, if the third standard data element set is matched only based on the characteristic word segmentation set of the target data element, it can also be combined with the first similarity and the corresponding third standard data elements in the third data element set. The first comprehensive score of each first candidate standard data element in the first candidate standardization list, from the third data element set and the first candidate standardization list, to determine at least one target standard data element for standardizing the target data element, so that no need Perform subsequent operations.

S309, based on the feature word segmentation set, determine a fourth data element set in the historical benchmarking record library that is similar to the target data element.

Wherein, the fourth data element set includes: at least one fourth standard data element whose feature information of the corresponding non-standard data element is highly similar to the feature word segmentation set, and the non-standard data element corresponding to the fourth standard data element. The second similarity between the feature information and the feature word segmentation set.

In this step S309, for each standard data element appearing in the historical benchmarking record library, the feature information of each non-standard data element of the corresponding standard data element can be regarded as a set, and then similar to the feature word segmentation set degree calculation, and finally determine at least one fourth standard data element whose feature information of the corresponding non-standard data element is highly similar to the feature word segmentation set.

In a possible implementation manner, in order to improve the speed of determining the fourth standard data element, this embodiment of the present application may also be based on at least one standard data element corresponding to the feature information of each non-standard data element in the historical benchmarking record library in advance , the constructed historical benchmarking record index library.

Wherein, the historical benchmarking record index database stores: a historical benchmarking index file corresponding to each standard data element appearing in the historical benchmarking record database. The historical benchmarking index file of each standard data element includes: based on at least one standard data element corresponding to the characteristic information of different non-standard data elements in the historical benchmarking database, each non-standard data element corresponding to the determined standard data element is determined. characteristic information.

Correspondingly, the fourth data element set in which the historical benchmarking index file is similar to the characteristic word segmentation set can be directly determined from the historical benchmarking index database, thereby improving the speed of retrieving the fourth data element set.

In this possible implementation manner, the fourth data element set includes: at least one fourth standard data element with a high similarity between the historical benchmarking index file and the feature word segmentation set, and a history of the fourth standard data element The second similarity between the benchmark index file and the feature word segmentation set.

S310: Determine the third data element set based on the first similarity corresponding to each third standard data element in the third data element set and the second similarity corresponding to each fourth standard data element in the fourth data element set and the second comprehensive score of each standard data element in the fourth data element set.

The process of determining the second comprehensive score may be similar to the process of determining the first comprehensive score.

For example, first, according to formula 2, the first similarity and the second similarity are respectively standardized. Secondly, the post-stimulation score corresponding to the first similarity of each standard data element and the post-stimulation score corresponding to the second similarity of each standard data element can be determined and calculated respectively according to formula 3. Finally, for each standard data element in the third data element set and the fourth data element set, the average value of the post-stimulation score of the first similarity and the post-stimulation score of the second similarity can be calculated. It is the second comprehensive score of the standard data element.

Among them, if a certain standard data element only belongs to the third data element set or the fourth data element set, then the post-stimulation score of the first similarity degree or the post-stimulation score of the second similarity degree can be regarded as zero, and on this basis , the second comprehensive score of the standard data element is actually the post-stimulation score of the first similarity or the post-stimulation score of the second similarity.

S311. Generate a second candidate normalization list for normalizing the target data element.

Wherein, the second candidate standardization list includes: at least one second candidate standard data element with a higher second comprehensive score determined from the third data element set and the fourth data element set, and the second candidate standard data Meta's second composite score.

In this embodiment, another possible standardized candidate list is determined through fuzzy matching through S307 to S311 to make up for the inability to match the candidate standard data elements or the poor matching effect through the precise matching of steps S302 to S306 The problem.

It can be understood that steps S302 to S306 can be executed synchronously with steps S307 to S311, or steps S302 to S306 can be executed first, and then steps S307 to S311 can be executed; or, steps S307 to S311 can be executed first and then executed Steps S302 to S306.

S312: Determine the first candidate standardization list and the second candidate standardization list by combining the first comprehensive score of each first candidate standard data element in the first candidate standardization list and the second comprehensive score of the second candidate standard data element in the second candidate standardization list The overall score for each candidate criterion data element in the standardized list.

It can be understood that, combining the first candidate standardization list and the second candidate standardization list whether each candidate standard data element appears in the two lists at the same time, and the first comprehensive score and the second comprehensive score of each candidate standard data element. At least one of the scores determines an overall score for the candidate standard data element. The overall score may reflect the integrated fitness of the candidate standard data elements for normalizing the target data elements.

Among them, there are various possibilities for determining the overall score.

For example, in a possible implementation manner, the weights of the first comprehensive score and the second comprehensive score may be set, and then, for each candidate standard data element in the first candidate standardization list and the second candidate standardization list, you may calculate The weighted summation of the first comprehensive score and the second comprehensive score of the candidate standard data element, and the result of the weighted summation is used as the overall score of the candidate standard data element.

Of course, if the candidate standard data element only belongs to the first candidate standardization list or the second candidate standardization list, then in the weighted summation, it can be considered that the value of the first comprehensive score or the second comprehensive score that the candidate standard data element does not have is the value of zero.

In yet another possible implementation manner, for each candidate normalization data element in the first candidate normalization list and the second candidate normalization list, it may be based on the feature information of the candidate normalization data element and the first integrated data element possessed by the candidate normalization data element At least one of the score and the second comprehensive score uses a pre-trained weight evaluation model to determine the first weight and the second weight corresponding to the candidate standard data element.

Among them, if the candidate standardized data element only has the first comprehensive score but not the second comprehensive score, it can be considered that its second comprehensive score is zero; for the case where the candidate standard data element only has the second comprehensive score, it can be considered that It has a first comprehensive score of zero, so that the first comprehensive score, the second comprehensive score and the reward and punishment score of the candidate standard data element are all input into the weight evaluation model to predict the two weights.

The first weight is the weight proportion of the first comprehensive score of the candidate standard data element, and the second weight is the weight proportion of the second comprehensive score of the candidate standard data element.

The weight evaluation model is obtained by training using the feature information corresponding to a plurality of data element samples marked with labels selected by the user, the first comprehensive score and the second comprehensive score. The data element sample may be a plurality of standard data elements recommended to the user determined in history, and the label of the data element sample indicates whether the data element is selected by the user as a standard data element.

The weight evaluation model may be obtained by training a random forest or an eXtreme Gradient Boosting (eXtreme Gradient Boosting, XGBOOST) model.

S313: Determine at least one target for normalizing the target data element from the first candidate normalization list and the second candidate normalization list in combination with the overall score of each candidate standard data element in the first candidate normalization list and the second candidate normalization list Standard data element.

For example, a set number of candidate standard data elements with higher overall scores may be selected as target standard data elements, so as to recommend the target standard data elements to the user for selection.

In this embodiment, on the basis of directly matching the standard data elements from the standard data element library and the historical benchmarking records in combination with the characteristic information of the target data elements to be standardized, the characteristic information of the target data elements is also combined into a The text is segmented, and based on the feature segment set composed of the segmented words from the text, fuzzy matching is performed on the standard data element library and the historical benchmarking records, so that the standard matching the target data element can be queried more comprehensively. The data element can also reduce the situation that the standard data element cannot be matched because the characteristic information based on the target data element is not covered by the standard data element library.

Moreover, combining the comprehensive scores of the standard data elements matched by these two methods to determine the target standard data element that is finally used to standardize the target data element is conducive to more comprehensive and accurate selection of the target standard data element, and improves the determined target. The standard data element is suitable for normalizing the fitness of the target data element.

It can be understood that a standard data element suitable for a standardized target data element generally belongs to the same entity category as the target data element. On this basis, the present application can also comprehensively consider whether the entity category of each candidate standard data element in the first candidate standardization list and the second candidate standardization list is consistent with the entity category of the target data element to determine the candidate standard data element. overall rating.

FIG. 5 shows a schematic flowchart of an implementation of determining the overall score of candidate standard data elements in the present application. The method of this embodiment may include:

S501 , based on at least one feature information of the target data element and a pre-built entity rule base, determine the entity category to which the target data element belongs.

The entity category to which the data element belongs can be used to represent the type of object guaranteed by the data element.

Feature matching rules that need to be satisfied by feature information of different entity categories are recorded in the entity rule base. If at least one feature information of the target data element matches the feature matching rule corresponding to a certain entity category in the entity rule base, then the target data element belongs to the entity category.

In an example, the form of data recorded in the entity rule base may be as shown in Table 1 below:

table 3

In Table 3, each entity name corresponds to an entity type (also called entity category), and a regular expression that needs to be satisfied by the feature information belonging to the entity type. For example, taking the ID card number as an example, the entity type of the ID card number is "person". If the feature information of the target data element to be standardized can match the regular expression corresponding to the ID card number, then the target data The entity type of the meta is "person".

S502: Determine the entity type of each candidate standard data element in the first candidate standardization list and the second candidate standardization list.

It can be understood that, if the entity type of the standard data element is stored in the standard data element library, the entity type corresponding to the candidate standard data element can be directly queried from the standard data element library.

If the entity type of the standard data element is not stored in the standard data element base, the entity type corresponding to the candidate standard data element can be matched from the entity rule base according to the feature information of the candidate standard data element.

S503, for each candidate standard data element in the first candidate standardization list and the second candidate standardization list, determine the reward and punishment score of the candidate standard data element.

Wherein, if the entity category of the candidate standard data element is the same as the entity category of the target data element, the reward and punishment score of the candidate standard data element is a positive number, such as a set positive score value. If the entity type of the candidate standard data element is different from the entity type of the target data element, the reward and punishment score of the candidate standard data element is a negative number, for example, the reward and punishment score is a set negative score value.

S504, determine the The overall score of each candidate standard data element in the first candidate standardization list and the second candidate standardization list.

For example, the overall score can be comprehensively determined by combining whether the candidate normalization data element belongs to two candidate normalization lists at the same time, its comprehensive score, and the value of the reward and punishment score. For example, for the candidate standard data elements contained in both the first candidate standardization list and the second candidate standardization list, the reward and punishment score of the candidate standard data element is a positive number, then the first comprehensive score and the second comprehensive score of the candidate standard data element are The higher the score, the higher the overall score for that candidate criterion data element.

In one example, for each candidate normalization data element in the first candidate normalization list and the second candidate normalization list, it may be based on the feature information of the candidate normalization data element and the first comprehensive score and the second comprehensive score of the candidate normalization data element. At least two of the score and the reward and punishment score are determined by using the pre-trained weight evaluation model to determine the first weight, the second weight and the third weight corresponding to the candidate standard data element.

Wherein, the first weight and the second weight respectively correspond to the weight ratio of the first comprehensive score and the second comprehensive score of the candidate standardized data element, as described above. The third weight is the weight ratio of the reward and punishment score of the candidate standard data element.

Similar to the above, the weight evaluation model is obtained by training the feature information, the first comprehensive score, the second comprehensive score and the reward and punishment score corresponding to multiple data element samples marked with labels selected by the user.

Correspondingly, for each candidate standardization data element in the first candidate standardization list and the second candidate standardization list, the first weight, the second weight and the third weight corresponding to the candidate standard data element are combined, and the candidate standard data element has The first comprehensive score, the second comprehensive score and the reward and punishment score are determined to determine the overall score of the candidate standard data element.

It can be understood that, after the overall scores of the candidate standard data elements are determined, at least one target standard data element for normalizing the target data elements can also be determined in combination with the overall scores of the candidate standard data elements.

In Fig. 5, considering whether the entity types of the candidate standard data element and the target data element are consistent, the overall score of the candidate standard data element is determined, so that the overall score can more accurately reflect the candidate standard data element as the standard data element of the target data element. The degree of suitability of the standard data element for standardizing the target data element can be more accurately determined.

In order to facilitate the understanding of the solution of the present application, the following description will be given with reference to the implementation block diagram of the solution of the present application. As shown in FIG. 6 , it shows a schematic diagram of an implementation principle framework of the method for determining a standard data element provided by an embodiment of the present application.

It can be seen from FIG. 6 that, after obtaining the feature information of the target data element to be standardized, the present application will perform fine matching and fuzzy matching respectively based on the feature information.

On the one hand, in the process of fine matching, based on the characteristic information of the target data element, the present application matches standard data elements similar to the characteristic information of the target data element from the standard data element library and the historical benchmarking record library respectively.

In this application, the at least one feature information of the target data element may all be one or more feature information input by the user, or may be multiple features mined on the basis of obtaining at least one feature information input by the user information.

For example, the feature information of the target data element that can be obtained by this application may at least include: at least one of the English code of the target data element and the Chinese name of the target data element.

The at least one characteristic information of the target data element may further include: a description of the target data element, a historical standard data element to which the target data element is historically normalized, and information of reference data elements of the same category as the target data element.

Wherein, the English code, Chinese name, description and information of the reference data element of the target data element can be input by the user.

The historical standard data element corresponding to the target data element may be input by the user, or may be queried from the historical standardization record on the user side according to the English code or Chinese name of the target data element.

Further, in order to improve the accuracy of standard data element matching, the present application can also obtain the feature information of the target data element, and analyze the code length label category, entity name and entity category of the target data element.

The code length label category of the target data element may be determined in combination with the length of each reference data element corresponding to the target data element.

For example, if the difference between the maximum length of each reference data element and the minimum length of each reference data element is zero, it can be determined that the code length tag type of the target data element is a fixed code.

If the length of each reference data element is less than or equal to the narrow code confidence level (for example, 10), the code length note type of the target data element is narrow code.

The entity name of the target data element may be matched from the entity rule base based on other feature information of the target data element; the entity name may also be identified by named entity recognition (NER) technology.

The entity category of the target data element can be obtained by querying the entity rule base, as described in the foregoing embodiments.

On the basis of the above, the present application can match the standard data elements with similar characteristic information from the standard data element database respectively according to the English code, Chinese name, historical standard data element information and entity type of the target data element, and obtain the standard data element The matching degree of the feature information of the target data element with the feature information of the target data element.

At the same time, according to the English code (or Chinese name) of the target data element, a non-standard data element whose English code matches the English code of the target data element can be matched from the historical benchmarking record database, and the non-standard data element in the history of the non-standard data element can be determined. Each standard data element to be normalized to and the number of normalization times corresponding to each standard data element. Then, the recommendation score of each standard data element can be determined based on the normalization times of each standard data element matched from the historical benchmarking records.

As shown in the exact matching part in Figure 6, combining one or both of the matching degree and recommendation score of each standard data element matched from the standard data element library and the historical benchmarking record library, the exact matching part can finally A first standardized list is determined, and the first standardized list includes the first comprehensive scores of the plurality of first candidate standardized data elements.

For the specific implementation of the precise matching in FIG. 6 , reference may be made to the related introduction of steps S302 to S306 in the embodiment of FIG. 3 above.

On the other hand, in the fuzzy matching process, the present application will construct a feature word segmentation set of the target data element based on the feature information of the target data element, search and match the standard data element index library and the historical benchmark record index library respectively, and combine The standard data elements obtained by searching the two index bases finally generate a second standard list. The second standardized list includes second comprehensive scores of the plurality of second candidate data elements.

On the other hand, for each candidate standard data element in the first standardized list and the second standardized list obtained by fuzzy matching and exact matching, the present application can also determine the reward and punishment of the candidate standard data element through the process shown in the embodiment of FIG. 5 Fraction.

As shown in Figure 6, if the candidate standard data element and the target data element belong to the same entity category, the reward and punishment score of the candidate standard data element is actually a reward score, and the reward score is a positive number; otherwise, the reward and punishment of the candidate standard data element The score is the penalty score, and the penalty score is a negative number.

On the basis of the above, this application combines the first comprehensive score, the second comprehensive score and the reward and punishment score of each candidate standard data element to carry out weighted score integration, and finally obtains the overall score of each candidate standard data element. See Figure 5 for details. The related introduction of step S504 in the embodiment.

Combined with the overall score of each candidate standard data element, at least one target standard data element suitable for recommendation as a standardization target data element can be determined, and the ordering of the at least one target standard data element can be determined.

In another aspect, corresponding to the method for determining a standard data element provided by the embodiments of the present application, the present application further provides an apparatus for determining a standard data element.

As shown in FIG. 7 , which shows a schematic structural diagram of a composition of an apparatus for determining a standard data element provided by an embodiment of the present application, the apparatus in this embodiment may include:

an information obtaining unit 701, configured to obtain at least one feature information of the target data element to be standardized;

A first set determining unit 702, configured to determine, based on each feature information of the target data element, a first data element set matching the target data element from a standard data element library, where the first data element set includes : each first standard data element matching the target data element in the standard data element library, and the degree of matching between the first standard data element and the target data element;

The second set determining unit 703 is configured to determine the second data element set corresponding to the target data element from the historical benchmarking record library based on the characteristic information of the target data element, and the historical benchmarking record library contains the second set of data elements. Stores: characteristic information of the non-standard data elements that have been standardized in the history, at least one standard data element that has been standardized in the history of the non-standard data elements, and the non-standard data elements that are historically normalized to each Standardization times of standard data elements; the second data element set includes: each second standard data element after the target data element has been standardized in history and the standardization times corresponding to the second standard data element;

The recommendation scoring unit 704 is configured to determine the recommendation score of the second standard data element according to the normalization times of the second standard data element in the second data element set, wherein the higher the normalization times of the second standard data element, the higher the recommendation score. more, the higher the recommendation score of the second standard data element is, and the recommendation score of the second standard data element is used to represent the suitability of the second standard data element as the standard data element of the target data element;

The first data element determining unit 705 is configured to combine the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set, from the first data element set. At least one first candidate standard data element for standardizing the target data element is determined in the data element set and the second data element set.

In a possible implementation manner, the historical benchmarking record library also stores: the non-standard data element was last normalized to the last normalization time of each standard data element in history;

Correspondingly, the second data element set determined by the second set determining unit further includes the latest standardized time corresponding to the second standard data element;

The recommended scoring unit includes:

A recommendation scoring subunit, configured to determine the recommendation score of the second standard data element according to the normalization times of the second standard data element in the second data element set and the last normalization time, wherein the second standard data element The more normalization times of the standard data element and the shorter the duration from the last normalization time to the current time, the higher the recommendation score of the second standard data element.

In yet another possible implementation, the first data element determining unit includes:

a first comprehensive scoring unit, configured to determine the first data element in combination with the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set the first comprehensive score of each standard data element in the set and the second data element set;

a first list generation unit, configured to generate a first candidate normalization list for normalizing the target data element, the first candidate normalization list includes: determining from the first data element set and the second data element set at least one first candidate standard data element with a higher first comprehensive score.

In another possible implementation manner, the first candidate standardization list generated by the first list generating unit further includes: a first comprehensive score of the first candidate standard data element;

The device also includes:

A word segmentation set construction unit, configured to construct a characteristic word segmentation set of the target data element based on at least one characteristic information of the target data element, the characteristic word set includes at least one characteristic word segmentation obtained from the word segmentation of the at least one characteristic information ;

A third set determining unit, configured to determine a third set of data elements in the standard data element library that is similar to the target data element based on the set of feature word segmentations, where the third set of data elements includes: a set of feature information and At least one third standard data element with a higher similarity of the feature word segment set, and the first similarity between the feature information set of the third standard data element and the feature word segment set; the feature information of the standard data element The set includes each characteristic information of the standard data element;

The target determination unit is configured to, based on the first similarity corresponding to each third standard data element and the first comprehensive score of each first candidate standard data element in the first candidate standardization list, from the third data element set and In the first candidate standardization list, at least one target standard data element for standardizing the target data element is determined.

In yet another possible implementation, the device further includes:

The fourth set determination unit is configured to determine, based on the feature word segmentation set, before the target determination unit determines at least one target standard data element used to standardize the target data element, determine the relationship between the target data element in the historical benchmarking record library and the target data element. A fourth data element set with similar data elements, the fourth data element set includes: at least one fourth standard data element whose feature information of the corresponding non-standard data element is highly similar to the feature word segmentation set and the The second similarity between the feature information of the non-standard data element corresponding to the fourth standard data element and the feature word segmentation set;

Correspondingly, the target determination unit includes:

The second comprehensive scoring unit is configured to be based on the first similarity corresponding to each third standard data element in the third data element set and the second similarity corresponding to each fourth standard data element in the fourth data element set , determining the second comprehensive score of each standard data element in the third data element set and the fourth data element set;

A second list generating unit, configured to generate a second candidate normalization list for normalizing the target data element, the second candidate normalization list includes: determining from the third data element set and the fourth data element set at least one second candidate standard data element with a higher second comprehensive score obtained, and the second comprehensive score of the second candidate standard data element;

The overall scoring unit is configured to combine the first comprehensive score of each first candidate standard data element in the first candidate standardization list and the second comprehensive score of the second candidate standard data element in the second candidate standardization list to determine the first candidate standard data element. The overall score of each candidate standard data element in the candidate standardization list and the second candidate standardization list;

A target comprehensive determination unit, configured to combine the overall scores of each candidate standard data element in the first candidate standardization list and the second candidate standardization list, and determine from the first candidate standardization list and the second candidate standardization list for standardization At least one target standard data element of the target data element.

In yet another possible implementation, the device further includes:

The category query unit is configured to, before the overall scoring unit determines the overall score of each candidate standard data element in the first candidate standardization list and the second candidate standardization list, based on at least one feature information of the target data element and a pre-built An entity rule base, which determines the entity category to which the target data element belongs, and records the feature matching rules that need to be satisfied by the feature information of different entity categories in the entity rule base;

a category determination unit, configured to determine the entity category of each candidate standard data element in the first candidate standardization list and the second candidate standardization list;

The reward and punishment score determination unit is used for, for each candidate standard data element in the first candidate standardization list and the second candidate standardization list, to determine the reward and punishment score of the candidate standard data element, wherein, if the candidate standard data element has a The entity category is the same as the entity category of the target data element, and the reward and punishment score of the candidate standard data element is a positive number; otherwise, the reward and punishment score of the candidate standard data element is a negative number;

Correspondingly, the overall scoring unit is specifically configured to combine the first comprehensive score and reward and punishment score of each first candidate standard data element in the first candidate standardization list, and the second candidate standard data in the second candidate standardization list. The second comprehensive score and reward and punishment score of the element are used to determine the overall score of each candidate standard data element in the first candidate standardization list and the second candidate standardization list.

In yet another possible implementation, the overall scoring unit includes

A weight determination subunit, configured to, for each candidate normalization data element in the first candidate normalization list and the second candidate normalization list, based on the feature information of the candidate normalization data element and the first normalization data element that the candidate normalization data element has At least two of the comprehensive score, the second comprehensive score, and the reward and punishment score, using a pre-trained weight evaluation model to determine the first weight, the second weight, and the third weight corresponding to the candidate standard data element;

an overall scoring subunit, configured to combine the first weight, second weight and third weight corresponding to the candidate standard data element for each candidate standardization data element in the first candidate standardization list and the second candidate standardization list, And the first comprehensive score, the second comprehensive score and the reward and punishment score that the candidate standard data element has, determine the overall score of the candidate standard data element;

Wherein, the first weight is the weight proportion of the first comprehensive score of the candidate standard data element, the second weight is the weight proportion of the second comprehensive score of the candidate standard data element, and the The third weight is the weight ratio of the reward and punishment score of the candidate standard data element;

The weight evaluation model is obtained by training using the feature information, the first comprehensive score, the second comprehensive score and the reward and punishment score corresponding to a plurality of data element samples marked with labels selected by the user.

In yet another possible implementation manner, the third set determining unit is specifically configured to, based on the data element index files corresponding to different standard data elements stored in the standard data element index library, select from the standard data element index library Determine a third data element set that is similar to the feature word segmentation set in the data element index file;

Wherein, the data element index file of the standard data element stores the characteristic information of the standard data element recorded in the standard data element library;

The third data element set includes: at least one third standard data element whose similarity between the data element index file and the feature word segmentation set is high, and the data element index file of the third standard data element is the same as that of the third standard data element. Describe the first similarity of the feature word set.

In yet another possible implementation manner, the fourth set determination unit is specifically configured to, based on the historical benchmarking index files corresponding to different standard data elements stored in the historical benchmarking record index database, select from the historical benchmarking index database Determine a fourth data element set that is similar to the feature word segmentation set in the historical benchmarking index file;

Wherein, the historical benchmarking index file corresponding to the standard data element includes: based on at least one standard data element corresponding to the characteristic information of different non-standard data elements in the historical benchmarking database, the determined standard data element corresponding to the standard data element Characteristic information of each non-standard data element;

The fourth data element set includes: at least one fourth standard data element with a high similarity between the historical benchmarking index file and the feature word segmentation set, and the historical benchmarking index file of the fourth standard data element and the The second similarity of the feature word segment set.

In another possible implementation manner, the word segmentation set construction unit includes:

a text combining unit for combining at least one feature information of the target data element into text;

The text segmentation unit is used for segmenting the text to obtain a feature segment set consisting of at least one feature segment obtained from the text segment.

In yet another possible implementation, the device further includes:

A feature cleaning unit, configured to, for each feature information of the target data element, based on a plurality of standard features stored in a standard feature library, before the text combining unit combines at least one feature information of the target data element into text, A standard feature matching the feature information is determined, and the feature information is standardized based on the matched standard feature to obtain at least one standardized feature information corresponding to the target data element.

The apparatus for determining standard data elements provided by the embodiments of the present application can be applied to computer equipment. The computer device may be an independent server, or may be a cloud platform or a server in a distributed system, or the like. Optionally, FIG. 8 shows a block diagram of the hardware structure of a computer device to which the solution of the present application is applicable. Referring to FIG. 8 , the hardware structure of the computer device may include: at least one processor 801 , at least one communication interface 802 , and at least one memory 803 and at least one communication bus 804;

In the embodiment of the present application, the number of the processor 801, the communication interface 802, the memory 803, and the communication bus 804 is at least one, and the processor 801, the communication interface 802, and the memory 803 complete the communication with each other through the communication bus 804;

The processor 801 may be a central processing unit CPU, or a specific integrated circuit ASIC (Application Specific Integrated Circuit), or one or more integrated circuits configured to implement the embodiments of the present invention, etc.;

The memory 803 may include high-speed RAM memory, and may also include non-volatile memory (non-volatile memory), etc., such as at least one disk memory;

Wherein, the memory stores a program, and the processor can call the program stored in the memory, and the program is used for:

Obtain at least one feature information of the target data element to be standardized;

Based on each feature information of the target data element, a first data element set matching the target data element is determined from the standard data element library, where the first data element set includes: Each first standard data element matched by the target data element, and the degree of matching between the first standard data element and the target data element;

Based on the characteristic information of the target data element, the second data element set corresponding to the target data element is determined from the historical benchmarking record library, and the historical benchmarking record library stores: feature information of the non-standard data element, at least one standard data element after the non-standard data element has been standardized in the history, and the non-standard data element has been standardized as the normalization times of each standard data element in the history; the The second data element set includes: each second standard data element that has been standardized in the history of the target data element and the normalization times corresponding to the second standard data element;

The recommendation score of the second standard data element is determined according to the normalization times of the second standard data element in the second data element set, wherein, the higher the normalization times of the second standard data element, the higher the second standard data element. The higher the recommendation score of the data element is, the recommendation score of the second standard data element is used to represent the suitability of the second standard data element as the standard data element of the target data element;

Combining the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set, from the first data element set and the second data element set At least one first candidate standard data element for normalizing the target data element is determined.

Optionally, the refinement function and extension function of the program may refer to the above description.

An embodiment of the present application further provides a storage medium, where the storage medium can store a program suitable for the processor to execute, and the program is used for:

Obtain at least one feature information of the target data element to be standardized;

Based on each feature information of the target data element, a first data element set matching the target data element is determined from the standard data element library, where the first data element set includes: Each first standard data element matched by the target data element, and the degree of matching between the first standard data element and the target data element;

Based on the characteristic information of the target data element, the second data element set corresponding to the target data element is determined from the historical benchmarking record library, and the historical benchmarking record library stores: feature information of the non-standard data element, at least one standard data element after the non-standard data element has been standardized in the history, and the non-standard data element has been standardized as the normalization times of each standard data element in the history; the The second data element set includes: each second standard data element that has been standardized in the history of the target data element and the normalization times corresponding to the second standard data element;

The recommendation score of the second standard data element is determined according to the normalization times of the second standard data element in the second data element set, wherein, the higher the normalization times of the second standard data element, the higher the second standard data element. The higher the recommendation score of the data element is, the recommendation score of the second standard data element is used to represent the suitability of the second standard data element as the standard data element of the target data element;

Combining the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set, from the first data element set and the second data element set At least one first candidate standard data element for normalizing the target data element is determined.

Optionally, the refinement function and extension function of the program may refer to the above description.

Finally, it should also be noted that in this document, relational terms such as first and second are used only to distinguish one entity or operation from another, and do not necessarily require or imply these entities or that there is any such actual relationship or sequence between operations. Moreover, the terms "comprising", "comprising" or any other variation thereof are intended to encompass a non-exclusive inclusion such that a process, method, article or device that includes a list of elements includes not only those elements, but also includes not explicitly listed or other elements inherent to such a process, method, article or apparatus. Without further limitation, an element qualified by the phrase "comprising a..." does not preclude the presence of additional identical elements in a process, method, article or apparatus that includes the element.

The various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The various embodiments can be combined as required, and the same and similar parts can be referred to each other. .

The above description of the disclosed embodiments enables any person skilled in the art to make or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, this application is not intended to be limited to the embodiments shown herein, but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (14)

1. a method for determining standard data element, is characterized in that, comprises: Obtain at least one feature information of the target data element to be standardized; Based on each feature information of the target data element, a first data element set matching the target data element is determined from the standard data element library, where the first data element set includes: Each first standard data element matched by the target data element, and the degree of matching between the first standard data element and the target data element; Based on the characteristic information of the target data element, the second data element set corresponding to the target data element is determined from the historical benchmarking record library, and the historical benchmarking record library stores: feature information of the non-standard data element, at least one standard data element after the non-standard data element has been standardized in the history, and the non-standard data element has been standardized as the normalization times of each standard data element in the history; the The second data element set includes: each second standard data element that has been standardized in the history of the target data element and the normalization times corresponding to the second standard data element; The recommendation score of the second standard data element is determined according to the normalization times of the second standard data element in the second data element set, wherein, the higher the normalization times of the second standard data element, the higher the second standard data element. The higher the recommendation score of the data element is, the recommendation score of the second standard data element is used to represent the suitability of the second standard data element as the standard data element of the target data element; Combining the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set, from the first data element set and the second data element set At least one first candidate standard data element for normalizing the target data element is determined. 2. The method according to claim 1, wherein the historical benchmarking record library also stores: the non-standard data element is standardized as the last standardized time of each standard data element in the history of the non-standard data element; The second data element set further includes: the latest standardized time corresponding to the second standard data element; The determining the recommendation score of the second standard data element according to the normalization times of the second standard data element in the second data element set includes: A recommendation score of the second standard data element is determined according to the normalization times of the second standard data element in the second data element set and the last normalization time, wherein the higher the normalization times of the second standard data element and the shorter the duration from the last normalized time to the current time, the higher the recommendation score of the second standard data element. 3 . The method according to claim 1 , wherein the matching degree of each first standard data element in the first data element set and the recommendation of each second standard data element in the second data element set are combined. 3 . Scoring, determining at least one first candidate standard data element for normalizing the target data element from the first data element set and the second data element set, including: Combining the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set, it is determined that the first data element set and the second data element set are in the The first comprehensive score of each standard data element; generating a first candidate normalization list for normalizing the target data element, where the first candidate normalization list includes: a first comprehensive score determined from the first data element set and the second data element set is higher of at least one first candidate standard data element. 4. The method according to claim 3, wherein the first candidate standardization list further comprises: a first comprehensive score of the first candidate standard data element; The method also includes: Based on at least one feature information of the target data element, construct a feature word segment set of the target data element, and the feature word set includes at least one feature word segment obtained from the at least one feature information segment; Based on the feature word segmentation set, determine a third data element set in the standard data element library that is similar to the target data element, where the third data element set includes: the similarity between the feature information set and the feature word segmentation set Higher at least one third standard data element, and the first similarity between the feature information set of the third standard data element and the feature word segmentation set; the feature information set of the standard data element includes the each characteristic information; Based on the first similarity corresponding to each third standard data element and the first comprehensive score of each first candidate standard data element in the first candidate standardization list, from the third data element set and the first candidate standardization In the list, at least one target standard data element for normalizing the target data element is determined. 5. The method according to claim 4, wherein before the determining at least one target standard data element for standardizing the target data element, further comprising: Based on the feature word segmentation set, a fourth data element set in the historical benchmarking record library that is similar to the target data element is determined, and the fourth data element set includes: the feature information of the corresponding non-standard data element and the The second similarity between the feature information of at least one fourth standard data element with a higher similarity of the feature word segment set and the non-standard data element corresponding to the fourth standard data element and the feature word segment set; Wherein, determining at least one target standard data element for standardizing the target data element includes: The third data element is determined based on the first similarity corresponding to each third standard data element in the third data element set and the second similarity corresponding to each fourth standard data element in the fourth data element set the second comprehensive score of each standard data element in the set and the fourth data element set; generating a second candidate normalization list for normalizing the target data element, the second candidate normalization list including: a second comprehensive score determined from the third data element set and the fourth data element set is higher at least one second candidate standard data element, and the second comprehensive score of the second candidate standard data element; Combining the first comprehensive score of each first candidate standard data element in the first candidate standardization list and the second comprehensive score of the second candidate standard data element in the second candidate standardization list, the first candidate standardization list and the first candidate standardization list are determined. 2. The overall score of each candidate standard data element in the candidate standardization list; In combination with the overall score of each candidate standard data element in the first candidate standardization list and the second candidate standardization list, determine from the first candidate standardization list and the second candidate standardization list for standardizing the target data element at least A target standard data element. 6. The method according to claim 5, wherein before the determining the overall score of each candidate standard data element in the first candidate standardization list and the second candidate standardization list, the method further comprises: Based on at least one feature information of the target data element and a pre-built entity rule base, determine the entity category to which the target data element belongs, and the entity rule base records the feature matching that needs to be satisfied by the feature information of different entity categories rule; determining the entity category of each candidate standard data element in the first candidate standardization list and the second candidate standardization list; For each candidate standard data element in the first candidate standardization list and the second candidate standardization list, determine the reward and punishment score of the candidate standard data element, wherein, if the entity category of the candidate standard data element is the same as the target data The entity category of the element is the same, and the reward and punishment score of the candidate standard data element is a positive number; otherwise, the reward and punishment score of the candidate standard data element is a negative number; Wherein, determining the overall score of each candidate standard data element in the first candidate standardization list and the second candidate standardization list includes: Determine the said The overall score of each candidate standard data element in the first candidate standardization list and the second candidate standardization list. 7. The method according to claim 6, wherein the determining the overall score of each candidate standard data element in the first candidate standardization list and the second candidate standardization list comprises the following steps: For each candidate normalization data element in the first candidate normalization list and the second candidate normalization list, based on the feature information of the candidate normalization data element and the first comprehensive score and the second comprehensive score possessed by the candidate normalization data element and at least two of the reward and punishment scores, using a pre-trained weight evaluation model to determine the first weight, second weight and third weight corresponding to the candidate standard data element; For each candidate standardization data element in the first candidate standardization list and the second candidate standardization list, combine the first weight, second weight and third weight corresponding to the candidate standard data element, and the candidate standard data element Having the first comprehensive score, the second comprehensive score and the reward and punishment score, determine the overall score of the candidate standard data element; Wherein, the first weight is the weight proportion of the first comprehensive score of the candidate standard data element, the second weight is the weight proportion of the second comprehensive score of the candidate standard data element, and the The third weight is the weight ratio of the reward and punishment score of the candidate standard data element; The weight evaluation model is obtained by training using the feature information, the first comprehensive score, the second comprehensive score and the reward and punishment score corresponding to a plurality of data element samples marked with labels selected by the user. 8. The method according to claim 4, wherein the determining a third data element set in the standard data element library that is similar to the target data element based on the characteristic word segmentation set comprises: Based on the data element index files corresponding to different standard data elements stored in the standard data element index library, determine from the standard data element index library a third data element set that is similar to the feature word segmentation set in the data element index file; Wherein, the data element index file of the standard data element stores the characteristic information of the standard data element recorded in the standard data element library; The third data element set includes: at least one third standard data element whose similarity between the data element index file and the feature word segmentation set is high, and the data element index file of the third standard data element is the same as that of the third standard data element. Describe the first similarity of the feature word set. 9. The method according to claim 5, wherein the determining a fourth data element set similar to the target data element in the historical benchmarking record library based on the characteristic word segmentation set, comprising: Based on the historical benchmarking index files corresponding to different standard data elements stored in the historical benchmarking record index library, determine from the historical benchmarking index library a fourth data element that is similar to the feature word segmentation set in the historical benchmarking index file gather; Wherein, the historical benchmarking index file corresponding to the standard data element includes: based on at least one standard data element corresponding to characteristic information of different non-standard data elements in the historical benchmarking database, the determined standard data element corresponding to the standard data element Characteristic information of each non-standard data element; The fourth data element set includes: at least one fourth standard data element with a high similarity between the historical benchmarking index file and the feature word segmentation set, and the historical benchmarking index file of the fourth standard data element and the The second similarity of the feature word segment set. 10. The method according to claim 4, wherein the constructing the feature word segmentation set of the target data element based on at least one feature information of the target data element comprises: combining at least one feature information of the target data element into text; The text is segmented to obtain a feature segment set consisting of at least one feature segment obtained from the text segment. 11. The method according to claim 10, wherein before the combining at least one characteristic information of the target data element into text, the method further comprises: For each feature information of the target data element, based on a plurality of standard features stored in the standard feature library, a standard feature matching the feature information is determined, and the feature information is standardized based on the matched standard feature to obtain the at least one standardized feature information corresponding to the target data element. 12. A device for determining standard data elements, characterized in that it comprises: an information obtaining unit for obtaining at least one feature information of the target data element to be standardized; a first set determining unit, configured to determine a first set of data elements matching the target data element from a standard data element library based on each feature information of the target data element, the first set of data elements comprising: Each first standard data element in the standard data element library that matches the target data element, and the degree of matching between the first standard data element and the target data element; The second set determining unit is configured to determine, based on each feature information of the target data element, a second data element set corresponding to the target data element from the historical benchmarking record library, and the historical benchmarking record library stores There are: characteristic information of the non-standard data elements that have been standardized in the history, at least one standard data element that has been standardized in the history of the non-standard data elements, and, the non-standard data elements have been standardized into each standard in the history. Standardization times of data elements; the second data element set includes: each second standard data element that has been standardized in the history of the target data element and the standardization times corresponding to the second standard data element; A recommendation scoring unit, configured to determine the recommendation score of the second standard data element according to the normalization times of the second standard data element in the second data element set, wherein the more the normalization times of the second standard data element , the higher the recommendation score of the second standard data element is, the recommendation score of the second standard data element is used to represent the suitability of the second standard data element as a standard data element of the target data element; The first data element determining unit is configured to combine the matching degree of each first standard data element in the first data element set and the recommendation score of each second standard data element in the second data element set, from the first data element set. At least one first candidate standard data element for normalizing the target data element is determined in the meta set and the second data element set. 13. A computer device, comprising: a memory and a processor; the memory for storing programs; The processor is configured to execute the program to implement each step of the method for determining a standard data element according to any one of claims 1 to 11. 14. A storage medium on which a computer program is stored, wherein when the computer program is executed by a processor, each of the methods for determining a standard data element according to any one of claims 1 to 11 is implemented. step.

Images