KR102019860B1 - Credit rating calculation apparatus based on user's network activity information and operating method thereof - Google Patents

Abstract

Disclosed are a credit rating grade calculating device based on network activity information of a user, and a method of operating the same. An apparatus for calculating a credit rating based on a user's network activity information and an operation method thereof according to the present invention take into consideration whether a user inputs personal information, whether a social network service (SNS) is active, and negative factors on the SNS. By objectively calculating the credit rating level for the user, it is possible to support a credit evaluation that reflects the user's actual life activity compared to the existing credit rating method using only financial information.

Description

CREDIT RATING CALCULATION APPARATUS BASED ON USER'S NETWORK ACTIVITY INFORMATION AND OPERATING METHOD THEREOF}

The present invention relates to an apparatus and a method of operating the apparatus for supporting a user to calculate a credit rating level for the user based on activity information on a social network service (SNS) of the user.

Recently, with the emergence of various types of loan products, the necessity to perform a simple and accurate credit rating for a user who wants a loan is increasing.

The existing credit rating method checks the financial information such as whether the user can use the existing loan products, whether the credit card is not overdue or whether the communication fee is overdue, and then checks whether the user has the ability to repay the loan based on the financial information. It was done by way of confirmation.

Since the elements used in the existing credit rating methods include personal information requiring high security, there is a problem that there is a risk such as disclosure of personal information.

In addition, the existing credit rating method is a way that users who want to use a loan visits a financial institution directly and consults with a loan product manager, which causes a lot of inconvenience for users who want to use a loan product. .

Recently, as the use of social network services (SNS) increases with the spread of the Internet, an environment for determining what activities are performed by each user on the web has been established.

If the user's activity on the SNS is analyzed, it is possible to grasp the user's personality and the like and predict the creditworthiness of the user.

Therefore, there is a need for a study of a new credit rating rating technique that allows a user to easily calculate a credit rating for a user based on activity information on the user's SNS.

An apparatus for calculating a credit rating based on a user's network activity information and an operation method thereof according to the present invention take into consideration whether a user inputs personal information, whether a social network service (SNS) is active, and negative factors on the SNS. By objectively calculating the credit rating level for the user, it is to support the credit rating reflecting the user's actual life activity compared to the existing credit rating method using only financial information.

According to an embodiment of the present invention, an apparatus for calculating a credit rating based on network activity information of a user stores and maintains a personal information score table in which different reliability calculation scores are recorded corresponding to a plurality of predetermined types of personal information. A personal information score table maintaining unit, a personal information request unit for transmitting a request for transmitting personal information of the user to the client terminal when a credit evaluation calculation request is received from the client terminal of the user, from the client terminal, to the user. And receiving at least one first personal information selectively input on the client terminal by the type of the at least one first personal information among the types of the plurality of personal information stored on the personal information score table. At least one of the corresponding records A reliability calculation score extracting unit for extracting a first reliability calculation score, an account information requesting unit requesting transmission of account information for a social network service (SNS) to which the user subscribes to the client terminal; When the account information for the at least one first SNS is received from the client terminal as the account information for the SNS to which the user subscribes, a predetermined activity for the number of account information for the at least one first SNS A network influence score calculator for calculating a network influence score by multiplying scores, and collecting text data existing on the web according to account information for the at least one first SNS based on account information for the at least one first SNS To each of the at least one first SNS after performing web crawling. An emotional score calculator for calculating an emotional score having a negative value based on a frequency of use of negative words identified from the collected text data, and based on the at least one first reliability calculation score, the network influence score, and the emotional score And calculating a credit rating level for the user and transmitting a credit rating rating for the user to the client terminal.

In addition, according to an embodiment of the present invention, the operation method of the credit rating rating calculation device based on the user's network activity information is a personal information score table in which different reliability calculation scores are recorded corresponding to each of a plurality of predetermined types of personal information. Storing and maintaining the request; when a credit evaluation calculation request is received from the client terminal of the user, transmitting a request for transmission of the personal information of the user to the client terminal; from the client terminal, the client by the user When at least one first personal information selectively input on the terminal is received, the at least one first personal information stored in the personal information score table corresponds to a type of the at least one first personal information, respectively. At least one first confidence level Extracting a number, requesting transmission of account information for the SNS to which the user subscribes to the client terminal, at least one first account information for the SNS to which the user subscribes from the client terminal; When the account information for the SNS is received, calculating a network influence score by multiplying a vitality score for the number of account information for the at least one first SNS, and calculating the account information for the at least one first SNS. A negative that is identified from text data collected for each of the at least one first SNS after performing a web crawl to collect text data existing on the web according to the account information for the at least one first SNS on the basis of Calculating an emotional score having a negative value based on the frequency of use of the word; And calculating a credit rating for the user based on the first reliability calculation score, the network influence score, and the emotional score, and then transmitting the credit rating for the user to the client terminal.

An apparatus for calculating a credit rating based on a user's network activity information and an operation method thereof according to the present invention take into consideration whether a user inputs personal information, whether a social network service (SNS) is active, and negative factors on the SNS. By objectively calculating the credit rating level for the user, it is possible to support a credit evaluation that reflects the user's actual life activity compared to the existing credit rating method using only financial information.

1 is a diagram illustrating a structure of a credit rating grade calculation device based on network activity information of a user according to an embodiment of the present invention.
2 is a flowchart illustrating a method of operating a credit rating grade calculation device based on network activity information of a user according to an embodiment of the present invention.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings. This description is not intended to limit the invention to the specific embodiments, it should be understood to include all changes, equivalents, and substitutes included in the spirit and scope of the invention. In describing the drawings, similar reference numerals are used for similar components, and unless otherwise defined, all terms used in the present specification, including technical or scientific terms, may be used in the art to which the present invention pertains. It has the same meaning as is commonly understood by someone who has it.

1 is a diagram illustrating a structure of a credit rating grade calculation device based on network activity information of a user according to an embodiment of the present invention.

Referring to FIG. 1, the credit rating rating calculation device 110 based on the user's network activity information according to an embodiment of the present invention may include a personal information score table holding unit 111, a personal information requesting unit 112, and a reliability calculation. The score extractor 113, the account information requester 114, the network influence score calculator 115, the emotion score calculator 116, and the credit rating grade calculator 117 are included.

The personal information score table holding unit 111 stores and maintains a personal information score table in which different reliability calculation scores are recorded corresponding to each of a plurality of predetermined types of personal information.

For example, information may be recorded in the personal information score table as shown in Table 1 below.

Type of personal information Confidence score name 10 points age 5 points address 10 points Loan purpose 5 points Loan amount 15 points Salary 20 points ... ...

When the personal information requesting unit 112 receives a credit evaluation calculation request from the user's client terminal 140, the personal information requesting unit 112 transmits a request for transmission of the user's personal information to the client terminal 140.

At this time, according to an embodiment of the present invention, when the personal information request unit 112 receives a credit evaluation calculation request from the client terminal 140, the client terminal 140 requests the transmission of the personal information of the user In the meantime, a query message for inquiring whether the user agrees to use the personal information may be transmitted.

If the user inputs at least one first personal information onto the client terminal 140 while agreeing to use the personal information through the client terminal 140, the client terminal 140 causes the at least one first individual. When the information is transmitted to the credit rating rating calculation device 110, the reliability calculation score extractor 113 receives the at least one first personal information from the client terminal 140 and is stored on the personal information score table. At least one first reliability calculation score recorded corresponding to the at least one type of the first personal information among the plurality of types of personal information is extracted.

For example, the reliability calculation score extractor 113 may refer to the personal information score table as shown in Table 1 when 'name, age, address' is received from the client terminal 140 as the at least one first personal information. '10 points, 5 points, 10 points', which are recorded in correspondence with 'name, age, address', may be extracted as the at least one first reliability calculation score.

In this way, when the at least one first reliability score is extracted, the account information requesting unit 114 is responsible for account information for the social network service (SNS) to which the user subscribes to the client terminal 140. Request the transmission of

When the network influence score calculator 115 receives account information for at least one first SNS from the client terminal 140 as account information for the SNS to which the user subscribes, the network impact score calculator 115 accounts for the at least one first SNS. A network influence score is calculated by multiplying a predetermined activity score for the number of information.

For example, the selected activity score is referred to as '20 points', and if the account information for 'five' SNS is received from the client terminal 140 as the account information for the at least one first SNS, the network influence. The score calculator 115 calculates a network influence score called '100 points' by multiplying the selected activity score by '20 points' for' 5 'which is the number of account information for the at least one first SNS. Can be.

The emotion score calculator 116 crawls a web for collecting text data existing on the web according to the account information for the at least one first SNS based on the account information for the at least one first SNS. After performing the operation, an emotional score having a negative value is calculated based on a frequency of use of negative words identified from the collected text data for each of the at least one first SNS.

At this time, according to an embodiment of the present invention, the emotional score calculator 116 is a word dictionary database 121, a negative word dictionary database 122, a crawler 123, a word extractor 124, a negative word selector 125 And the emotional score calculation processor 126.

In the word dictionary database 121, a plurality of words and a similar word for each of the plurality of words are pre-allocated and stored in a vector corresponding to each other so that the Euclidean distance between the two words is smaller.

In this regard, information may be stored in the word dictionary database 121 as shown in Table 2 below.

Multiple words Attribute vector computer (1 2 3 4 5) depressed (6, 7, 8, 9, 10) revenge (11, 12, 13, 14, 15) cellphone (16, 17, 18, 19, 20) ... ...

Euclidean distance means the distance between two vectors and can be calculated according to Equation 1 below.

Where D is the Euclidean distance and p _i and q _i are the i th components included in the two vectors. The smaller the Euclidean distance between two vectors, the two vectors can be regarded as similar vectors, and the larger the Euclidean distance between the two vectors, the two vectors can be regarded as dissimilar vectors.

In relation to this, as shown in Table 2, the word dictionary database 121 calculates a smaller Euclidean distance as shown in Equation 1 according to the similarity between the words for each of the plurality of words by the administrator. The characteristic vector may be pre-allocated and stored.

The negative dictionary database 122 selects and stores only a plurality of negative words predetermined as negative words among the plurality of words.

For example, the negative dictionary database 122 may store information as shown in Table 3 below.

Multiple negative words depressed revenge financial obligation Overdue ...

The crawling unit 123 performs web crawling to collect text data existing on the web according to the account information for the at least one first SNS based on the account information for the at least one first SNS. The text data corresponding to each of the at least one first SNS is collected.

The word extractor 124 extracts a plurality of first words from text data corresponding to each of the at least one first SNS by performing a morpheme analysis on text data corresponding to each of the at least one first SNS. do.

In relation to this, when the at least one first SNS is 'SNS 1', 'SNS 2', 'SNS 3', 'SNS 4', and 'SNS 5', the crawler 123 may include 'SNS 1', Based on the account information for each of 'SNS 2', 'SNS 3', 'SNS 4', and 'SNS 5', 'SNS 1', 'SNS 2', 'SNS 3', 'SNS 4' and 'SNS 5 '' Performs web crawling to collect text data on the web according to account information for each of them, so that 'SNS 1', 'SNS 2', 'SNS 3', 'SNS 4', and 'SNS 5' Corresponding text data can be collected.

Then, the word extractor 124 performs morphological analysis on text data corresponding to 'SNS 1', 'SNS 2', 'SNS 3', 'SNS 4', and 'SNS 5', respectively. The plurality of first words may be extracted by extracting words from text data corresponding to 'SNS 2', 'SNS 3', 'SNS 4', and 'SNS 5'.

The negative word selector 125 checks a feature vector for each of the plurality of first words with reference to the word dictionary database 121 and on the negative word dictionary database 122 for each of the plurality of first words. By calculating the Euclidean distance between the characteristic vector of each of the stored plurality of negative words, the Euclidean distance calculated in the relationship with the plurality of negative words among the plurality of first words is within a predetermined first reference distance. Words that have been computed are selected as a plurality of first negative words.

For example, as a result of extracting the plurality of first words from 'SNS 1', 'SNS 2', 'SNS 3', 'SNS 4', and 'SNS 5', the word 'depression, computer' in 'SNS 1' The word 'debt, mobile phone' was extracted from 'SNS 2', the word 'patent and trademark' was extracted from 'SNS 3', the word 'economic' was extracted from 'SNS 4', When the word 'due fee' is extracted from the SNS 5, the negative word selector 125 refers to the word dictionary database 121 to each of 'depression, computer, debt, mobile phone, patent, trademark, economy, late fee'. We can check the characteristic vector for.

Then, the negative selector 125 compares each of the plurality of negative words stored in the negative dictionary database 122 for each of 'depression, computer, debt, mobile phone, patent, trademark, economy, late fee'. Euclidean distance between feature vectors can be calculated.

In this regard, the negative selector 125 may calculate a Euclidean distance between 'depression' and a feature vector of each of the plurality of negative words, and between the computer and the feature vector of each of the plurality of negative words. Euclidean distance can be computed, and Euclidean distance between 'debt' and a characteristic vector of each of the plurality of negative words can be computed, and Euclidean distance between 'mobile phone' and a feature vector of each of the plurality of negative words Compute a Euclidean distance between a 'patent' and a feature vector of each of the plurality of negative words, and calculate a Euclidean distance between a 'trademark' and a feature vector of each of the plurality of negative words Can calculate the Euclidean distance between the 'economy' and the characteristic vector of each of the plurality of negative words, and can be used to calculate the late fee and the plurality of negative words. It is possible to calculate the Euclidian distance between the properties of the respective vector.

Thereafter, the negative word selector 125 has a Euclidean distance calculated in relation to the plurality of negative words among 'depression, computer, debt, mobile phone, patent, trademark, economy, and late fees' within a predetermined first reference distance. Words that have been computed may be selected as the plurality of first negative words.

In this regard, if there is a case where the Euclidean distance computed in relation to the plurality of negative words for 'depression' is calculated within the predetermined first reference distance, 'depression' is one of the plurality of first negative words. Can be selected.

When the plurality of first negative words are selected, the emotional score calculation processor 126 multiplies a predetermined emotional reference score having a negative value with respect to the number of the plurality of first negative words to process the calculation of the emotional score.

For example, 'depression, debt, late fees' was selected as the plurality of first negative words among 'depression, computer, debt, mobile phone, patent, trademark, economy, and late fees', and the selected emotional reference score was '-2 points'. ', The emotional score calculation processing unit 126 multiplies' -3' which is the number of the plurality of first negative words by '-2' which is the selected emotional reference score to multiply the emotional score of '-6 points'. Can be calculated.

At this time, according to an embodiment of the present invention, the emotional score calculator 116 may further include a weight table holding unit 127, a type checking unit 128, a diffusion degree calculating unit 129 and a weight extracting unit 130. Can be.

The weight table maintaining unit 127 stores and maintains a weight table in which different emotion score correction weights are recorded corresponding to each of a plurality of predetermined diffusion ranges.

For example, information may be recorded in the weight table as shown in Table 4 below.

Multiple diffusion ranges Emotion score correction weight 0 ~ 0.1 One 0.1-0.2 1.1 0.2 ~ 0.3 1.12 ... ...

When the plurality of first negative words are selected, the type checking unit 128 checks the type of SNS that is a collection source of text data for each of the plurality of first negative words among the at least one first SNS.

In relation to this, when it is said that 'depression, debt, late fees' are selected as the plurality of first negative words as described above, the type checking unit 128 is a source of collecting text data for the word 'depression'. You can identify 'SNS 1' as the type of SNS, 'SNS 2' as the type of SNS that is the source of collection of text data about 'debt', and the text data for the word 'due fee'. SNS 5 can be identified as the type of SNS that is the source of collection.

When the type of SNS for each of the plurality of first negative words is confirmed, a diffusivity calculator 129 may determine that the plurality of first negative words are based on a classification ratio of the negative words for each of the at least one first SNS. A first diffusivity indicating how evenly distributed on one first SNS is calculated.

At this time, according to an embodiment of the present invention, the diffusion calculator 129 may calculate the first diffusion based on Equation 2 below.

Here, I is the first diffusivity, p _i is the ratio of the negative words identified and classified as the i-th SNS through the type identification unit 128 in the plurality of first negative words, J is the at least one first SNS Means the total number of.

The first diffusivity calculated according to Equation 2 is a measure indicating whether or not the plurality of first negative words are distributed to the at least one first SNS, and the value of the first diffusivity calculated according to Equation 2 below. The closer to 0, the more negative words are concentrated on a specific SNS among the at least one first SNS, and the closer to the value of the first diffusivity, the closer to 1, the at least one first SNS. This means that the plurality of first negative words are evenly distributed over the entire SNS.

In relation to this, according to the above-described example, 'depression, debt, late fees' is selected as the plurality of first negative words, and 'depression' is a negative word that is identified and classified as 'SNS 1' through the type checker 128 Assume that 'debt' is a negative that is identified and classified as 'SNS 2' and 'due fees' is a negative that is identified and classified as 'SNS 5'.

In this case, p ₁ is a ratio of negative words identified and classified as 'SNS 1' in the plurality of first negative words, and thus '1/3', and p ₂ is 'SNS 2' in the plurality of first negative words. Since it is the ratio of the negative to be identified and classified as '1/3', p ₃ is the ratio of the negative to be identified and classified as 'SNS 3' in the plurality of first negatives, since no negative is classified. , '0', and p ₄ is a ratio of negative words identified and classified as 'SNS 4' in the plurality of first negative words, and since no negative words are classified, p ₄ is '0' and p ₅ is the above. Since the ratio of the negative words identified and classified as 'SNS 5' in the plurality of first negative words is '1/3'.

In consideration of this point, when the first diffusion degree according to Equation 2 is calculated, the first diffusion degree may be calculated as '0.67'.

When the first spreading degree is calculated, the weight extracting unit 130 extracts a first emotional score correction weight recorded corresponding to the spreading range to which the first spreading degree belongs by referring to the weight table.

In this case, when the first emotional score correction weight is extracted, the emotional score calculation processor 126 multiplies the selected emotional reference score by the first emotional score correction weight to perform correction on the selected emotional reference score. The calculation of the emotional score may be processed by multiplying the corrected emotional reference score by the number of the plurality of first negative words.

In relation to this, if the first emotional score correction weight is extracted as '1.16' and the selected emotional reference score is '-2', the emotional score calculating unit 126 may use the selected emotional reference score '-' 2 points' may be multiplied by the first emotion score correction weight '1.16' to correct the selected emotion reference score as' -2.32 '.

Usually, when negative words are spread on various SNSs rather than only on specific SNSs, the psychological state of the user can be regarded as more unstable, so that the administrator can see on the weight table as shown in Table 4 above. The lower the diffusivity, the lower the emotional correction weight is recorded in correspondence, and the higher the diffusivity, the higher the emotional correction weight is recorded in correspondence. can do.

As such, when the calculation of the at least one first reliability calculation score, the network influence score, and the emotion score is completed, the credit rating grade calculation unit 117 may calculate the at least one first reliability calculation score and the network influence score. And after calculating the credit rating for the user based on the emotional score and transmits a credit rating for the user to the client terminal 140.

At this time, according to an embodiment of the present invention, the credit rating grade calculation unit 117 may include a credit rating table table holding unit 118, a sum score calculation unit 119 and a credit rating grade determination unit 120. .

The credit rating table holding unit 118 stores and maintains a credit rating table in which different credit rating levels are recorded corresponding to each of a plurality of predetermined score ranges.

For example, information may be recorded in the credit rating table as shown in Table 5 below.

Multiple score ranges Credit rating 0-30 points Grade 5 30-50 points Grade 4 50 to 80 points Grade 3 ... ...

When the at least one first reliability calculation score, the network influence score, and the emotion score are calculated, an addition score calculator 119 may add the at least one first reliability calculation score, the network influence score, and the emotion score. Calculate the total score.

The credit rating rating determination unit 120 extracts a first credit rating rating recorded corresponding to the score range to which the sum score belongs, by referring to the credit rating rating table and sets the first credit rating rating to the user. Determined by credit rating.

In this regard, the at least one first confidence score was calculated as '10, 5, 10 ', the network influence score was calculated as' 100,' and the emotional score was' -50 '. If it is calculated, the sum score calculation unit 119 may calculate a sum score of '75 points ', and the credit rating grade determining unit 120 refers to the credit score table as shown in Table 5 above, and scores '75 points'. '3 ratings', which are credit ratings recorded corresponding to '3', may be determined as a credit rating level for the user.

As a result, the credit rating rating calculation device 110 based on the user's network activity information according to the present invention determines the credit rating rating for the user in consideration of whether the user's input of personal information, whether or not the SNS activity and negative factors on the SNS. By calculating objectively, it is possible to support a credit evaluation that reflects a user's actual life activity compared to a conventional credit rating method using only financial information.

According to an embodiment of the present invention, the credit rating rating calculation device 110 based on the user's network activity information 110 is the credit rating rating based on the at least one first reliability calculation score, the network influence score and the emotion score In addition to calculating the, the positive activity on the SNS based on the additional score or the SNS account information of the user according to the verification result of the authenticity of the at least one first personal information received from the client terminal 140 It may be configured to calculate the credit rating rating in consideration of the index, the psychological test analysis index, the additional score in consideration of the response or response by other users, and the like.

2 is a flowchart illustrating a method of operating a credit rating grade calculation device based on network activity information of a user according to an embodiment of the present invention.

In step S210, a personal information score table in which different reliability calculation scores are corresponded and recorded for each type of a plurality of predetermined personal information is stored and maintained.

In step S220, when a credit evaluation calculation request is received from the user's client terminal, the client terminal transmits a request for transmission of personal information of the user.

In step S230, when the at least one first personal information selected and input by the user on the client terminal is received from the client terminal, the type of the plurality of personal information stored on the personal information score table. And extracting at least one first reliability calculation score recorded corresponding to each of the at least one type of first personal information.

In step S240, the client terminal requests the transmission of account information for the SNS to which the user subscribes.

In step S250, when the account information for the at least one first SNS is received from the client terminal as account information for the SNS to which the user subscribes, the number of account information for the at least one first SNS is determined. The network influence score is calculated by multiplying the selected activity score.

In step S260, after performing a web crawl for collecting text data existing on the web according to the account information for the at least one first SNS based on the account information for the at least one first SNS, An emotional score having a negative value is calculated based on the frequency of use of negative words identified from the collected text data for each of the at least one first SNS.

In step S270, after calculating a credit rating for the user based on the at least one first reliability score, the network influence score, and the emotion score, the credit rating for the user is transmitted to the client terminal. do.

At this time, according to an embodiment of the present invention, in step S270, storing and maintaining a credit rating table, in which a different credit rating rating is recorded corresponding to each of a plurality of predetermined score ranges, wherein the at least Calculating a sum score by adding up the at least one first reliability score, the network influence score, and the emotion score, when one first reliability score, the network influence score, and the emotion score are calculated; And extracting a first credit rating grade recorded corresponding to the score range to which the sum score belongs by referring to the rating grade table and determining the first credit rating rating as a credit rating rating for the user. have.

Further, according to an embodiment of the present invention, in step S260, the pre-assigned feature vectors correspond to each other such that a plurality of words and similar words for each of the plurality of words are calculated to have a smaller Euclidean distance between the two words. Maintaining a stored word dictionary database; maintaining a negative word dictionary database in which only a plurality of negative words predetermined as a negative word among the plurality of words are selected and stored; and in the at least one first SNS Collects text data corresponding to each of the at least one first SNS by performing a web crawl to collect text data existing on the web according to the account information of the at least one first SNS based on the account information of the at least one first SNS. And text data corresponding to each of the at least one first SNS. Extracting a plurality of first words from text data corresponding to each of the at least one first SNS by identifying a feature vector for each of the plurality of first words by referring to the word dictionary database. And calculate a Euclidean distance between the feature vectors of each of the plurality of negative words stored in the negative dictionary database for each of the plurality of first words, thereby calculating the plurality of negative ones of the plurality of first words. Selecting words having a case where the Euclidean distance calculated in relation to the words is calculated within a predetermined first reference distance as a plurality of first negative words and a negative value for the number of the first negative words And multiplying the selected emotional reference score to process the calculation of the emotional score.

At this time, according to an embodiment of the present invention, in step S260, storing and maintaining a weight table in which different emotion score correction weights are correspondingly recorded for each of a plurality of predetermined diffusion ranges, and storing the plurality of weight score tables. When the first negative words of the selected ones, the type of the SNS that is the source of the collection of text data for each of the plurality of first negative words of the at least one first SNS is confirmed, each of the plurality of first negative words When the type of the SNS is determined, indicating how uniformly the plurality of first negative words are distributed on the at least one first SNS based on a classification ratio of the negative words for each of the at least one first SNS. Calculating a first diffusivity and, when the first diffusivity is computed, the spread to which the first diffusivity belongs with reference to the weight table The method may further include extracting a first emotion score correction weight recorded corresponding to the range.

In the processing of the emotional score, when the first emotional score correction weight is extracted, the selected emotional reference score is multiplied by the first emotional score correction weight to perform correction on the selected emotional reference score. Thereafter, the calculation of the emotional score may be processed by multiplying the corrected emotional reference score by the number of the plurality of first negative words.

At this time, according to an embodiment of the present invention, the step of calculating the first diffusivity may be calculated based on the equation (1).

The operating method of the credit rating grade calculation device based on the network activity information of the user according to an embodiment of the present invention has been described above with reference to FIG. 2. Here, the operation method of the credit rating rating calculation device based on the user's network activity information according to an embodiment of the present invention is based on the operation of the credit rating rating calculation device 110 based on the user's network activity information described with reference to FIG. Since it may correspond to the configuration for a more detailed description thereof will be omitted.

According to an embodiment of the present invention, an operation method of an apparatus for calculating a credit rating based on network activity information of a user may be implemented as a computer program stored in a storage medium for execution by combining with a computer.

In addition, the operating method of the credit rating rating calculation device based on the user's network activity information according to an embodiment of the present invention may be implemented in the form of program instructions that can be executed by various computer means may be recorded on a computer readable medium. . The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. Program instructions recorded on the media may be those specially designed and constructed for the purposes of the present invention, or they may be of the kind well-known and available to those having skill in the computer software arts. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tape, optical media such as CD-ROMs, DVDs, and magnetic disks, such as floppy disks. Magneto-optical media, and hardware devices specifically configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine code generated by a compiler, but also high-level language code that can be executed by a computer using an interpreter or the like.

In the present invention as described above has been described by the specific embodiments, such as specific components and limited embodiments and drawings, but this is provided to help a more general understanding of the present invention, the present invention is not limited to the above embodiments. For those skilled in the art, various modifications and variations are possible from these descriptions.

Therefore, the spirit of the present invention should not be limited to the described embodiments, and all the things that are equivalent to or equivalent to the claims as well as the following claims will belong to the scope of the present invention. .

110: credit rating rating device based on the user's network activity information
111: personal information score table holding unit 112: personal information request unit
113: confidence calculation score extractor 114: account information request unit
115: network influence score calculator 116: emotion score calculator
117: credit rating grade calculation unit 118: credit rating grade table maintenance unit
119: sum score calculation unit 120: credit evaluation grade determination unit
121: Word Dictionary Database 122: Negative Dictionary Database
123: crawling unit 124: word extraction unit
125: negative word selection unit 126: emotion score calculation processing unit
127: weight table holding unit 128: type checking unit
129: diffusivity calculator 130: weight extractor
140: client terminal

Claims (12)

A personal information score table holding unit for storing and maintaining a personal information score table in which different reliability calculation scores are correspondingly recorded according to a plurality of predetermined types of personal information;
A personal information request unit which transmits a request for transmission of personal information of the user to the client terminal when a credit evaluation calculation request is received from the user's client terminal;
When the at least one first personal information selected and input by the user on the client terminal is received from the client terminal, the at least one of the types of the plurality of personal information stored on the personal information score table. A reliability calculation score extracting unit for extracting at least one first reliability calculation score recorded corresponding to each type of first personal information;
An account information request unit for requesting transmission of account information for a social network service (SNS) to which the user subscribes to the client terminal;
When the account information for the at least one first SNS is received from the client terminal as account information for the SNS to which the user subscribes, the account information is predetermined for the number of account information for the at least one first SNS. A network influence score calculator for multiplying activity scores to calculate a network influence score;
The at least one after performing web crawling to collect text data existing on the web according to the account information for the at least one first SNS based on the account information for the at least one first SNS; An emotion score calculator configured to calculate an emotion score having a negative value based on a frequency of use of negative words identified from collected text data for each of the first SNSs of the first SNS; And
Calculating a credit rating level for the user based on the at least one first reliability score, the network influence score, and the emotion score, and calculating a credit rating level for transmitting the credit rating rating for the user to the client terminal part
Including,
The emotional score calculation unit
A word dictionary database, in which a plurality of words and a similar word for each of the plurality of words are stored in correspondence with each other, wherein a pre-allocated feature vector is stored so that an Euclidean distance between two words is calculated smaller;
A negative word dictionary database in which only a plurality of negative words predetermined as a negative word among the plurality of words are selected and stored;
A weight table holding unit for storing and maintaining a weight table in which different emotion score correction weights are correspondingly recorded for each of a plurality of predetermined diffusion ranges;
The web crawl for collecting text data existing on the web according to the account information for the at least one first SNS, based on the account information for the at least one first SNS, to perform the web crawling; A crawling unit for collecting text data corresponding to each;
A word extracting unit extracting a plurality of first words from text data corresponding to each of the at least one first SNS by performing morphological analysis on text data corresponding to each of the at least one first SNS;
Identify a feature vector for each of the plurality of first words with reference to the word dictionary database, and for each of the plurality of first words, for each of the plurality of negative words stored on the negative dictionary database; By calculating the Euclidean distance between the feature vectors of the plurality of first words, the Euclidean distance calculated in relation to the plurality of negative words is calculated within a predetermined first reference distance a plurality of words A negative word selection unit for selecting one negative word;
A type checking unit for checking a type of SNS which is a collection source of text data for each of the plurality of first negative words among the at least one first SNS when the plurality of first negative words are selected;
When the type of SNS for each of the plurality of first negative words is confirmed, the plurality of first negative words are displayed on the at least one first SNS based on a classification ratio of the negative words for each of the at least one first SNS. A diffusivity calculator for calculating a first diffusivity indicating how evenly distributed it is;
A weight extraction unit configured to extract a first emotion score correction weight recorded corresponding to the diffusion range to which the first diffusion belongs, when the first diffusion is calculated; And
When the first emotional score correction weight is extracted, the first emotional score score is multiplied by a predetermined emotional reference score having a negative value to perform the correction on the selected emotional reference score, and then the plurality of first negative words Emotion score calculation processing unit for processing the calculation for the emotional score by multiplying the corrected emotional reference score with respect to the number of
Including;
The diffusion calculation unit
A credit rating rating calculation device based on network activity information of a user calculating the first spreading degree based on Equation 1 below.
[Equation 1]

Here, I is the first diffusivity, p _i is the ratio of the negative words identified and classified as the i-th SNS through the type checking unit in the plurality of first negative words, J is the total number of the at least one first SNS Means. The method of claim 1,
The credit rating rating calculation unit
A credit rating table table holding unit for storing and maintaining a credit rating table in which different credit rating levels are recorded corresponding to each of a plurality of predetermined score ranges;
When the at least one first reliability calculation score, the network influence score, and the emotion score are calculated, an addition score that calculates a sum score by adding up the at least one first reliability calculation score, the network influence score, and the emotion score. A calculator; And
A credit rating rating for extracting a first credit rating that is recorded corresponding to a score range to which the sum score belongs, is determined by referring to the credit rating rating table to determine the first credit rating as a credit rating for the user. Decision
Credit rating rating calculation device based on the user's network activity information comprising a. delete delete delete Storing and maintaining a personal information score table in which different reliability calculation scores are correspondingly recorded according to a plurality of predetermined types of personal information;
If a credit evaluation calculation request is received from a user's client terminal, transmitting a request for transmission of personal information of the user to the client terminal;
When the at least one first personal information selected and input by the user on the client terminal is received from the client terminal, the at least one of the types of the plurality of personal information stored on the personal information score table. Extracting at least one first reliability calculation score recorded corresponding to each type of first personal information;
Requesting the client terminal to transmit account information for a social network service (SNS) to which the user subscribes;
When the account information for the at least one first SNS is received from the client terminal as account information for the SNS to which the user subscribes, the account information is predetermined for the number of account information for the at least one first SNS. Calculating a network influence score by multiplying a vitality score;
The at least one after performing web crawling to collect text data existing on the web according to the account information for the at least one first SNS based on the account information for the at least one first SNS; Calculating an emotional score having a negative value based on a frequency of use of negative words identified from the collected text data for each of the first SNSs of; And
Calculating a credit rating for the user based on the at least one first reliability score, the network influence score, and the emotion score, and then transmitting the credit rating for the user to the client terminal
Including,
Computing the emotional score
Maintaining a word dictionary database in which a plurality of words and a similar word for each of the plurality of words are pre-allocated with feature vectors pre-assigned so that Euclidean distance between the two words is calculated smaller;
Maintaining a negative dictionary database in which only a plurality of negative words predetermined as predetermined negative words among the plurality of words are selected and stored;
Storing and maintaining a weight table in which different emotion score correction weights are correspondingly recorded for each of a plurality of predetermined spreading ranges;
The web crawl for collecting text data existing on the web according to the account information for the at least one first SNS, based on the account information for the at least one first SNS, to perform the web crawling; Collecting text data corresponding to each;
Extracting a plurality of first words from text data corresponding to each of the at least one first SNS by performing morphological analysis on text data corresponding to each of the at least one first SNS;
Identify a feature vector for each of the plurality of first words with reference to the word dictionary database, and for each of the plurality of first words, for each of the plurality of negative words stored on the negative dictionary database; By calculating the Euclidean distance between the feature vectors of the plurality of first words, the Euclidean distance calculated in relation to the plurality of negative words is calculated within a predetermined first reference distance a plurality of words Selecting with 1 negatives;
If the plurality of first negative words are selected, identifying a type of SNS that is a collection source of text data for each of the plurality of first negative words among the at least one first SNS;
When the type of SNS for each of the plurality of first negative words is confirmed, the plurality of first negative words are displayed on the at least one first SNS based on a classification ratio of the negative words for each of the at least one first SNS. Calculating a first diffusivity indicating how evenly distributed it is;
Extracting a first emotion score correction weight recorded corresponding to a diffusion range to which the first diffusion belongs, by referring to the weight table when the first diffusion is calculated; And
When the first emotional score correction weight is extracted, the first emotional score is multiplied by a predetermined emotional reference score having a negative value to perform correction for the selected emotional reference score, and then the plurality of first negative words Processing the emotion score by multiplying the corrected emotion reference score by the number of eyes;
Including;
Computing the first diffusion degree
A method of operating a credit rating rating calculation device based on network activity information of a user calculating the first spreading degree based on Equation 2 below.
[Equation 2]

Here, I is the first spreading degree, p _i is the ratio of the negative words identified and classified as the i-th SNS through the step of identifying the type of the SNS in the plurality of first negative words, J is the at least one first 1 means total number of SNS. The method of claim 6,
The step of transmitting to the client terminal
Storing and maintaining a credit rating table in which different credit rating levels are correspondingly recorded for each of a plurality of predetermined score ranges;
Calculating the sum score by adding up the at least one first confidence score, the network influence score, and the emotion score when the at least one first confidence score, the network influence score, and the emotion score are calculated; And
Determining a first credit rating rating as a credit rating rating for the user by extracting a first credit rating rating recorded corresponding to a score range to which the sum score belongs by referring to the credit rating rating table;
Operating method of the credit rating rating calculation device based on the user's network activity information comprising a. delete delete delete A computer-readable recording medium having recorded thereon a program for performing the method of claim 6. A computer program stored in a storage medium for executing the method of any one of claims 6 or 7 in combination with a computer.

Images