CN117633187A - Method, device, equipment and storage medium for determining reply sentence - Google Patents
Method, device, equipment and storage medium for determining reply sentence Download PDFInfo
- Publication number
- CN117633187A CN117633187A CN202311661659.7A CN202311661659A CN117633187A CN 117633187 A CN117633187 A CN 117633187A CN 202311661659 A CN202311661659 A CN 202311661659A CN 117633187 A CN117633187 A CN 117633187A
- Authority
- CN
- China
- Prior art keywords
- description
- processed
- sentence
- problem description
- sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/332—Query formulation
- G06F16/3329—Natural language query formulation
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Prevention of errors by analysis, debugging or testing of software
- G06F11/3668—Testing of software
- G06F11/3672—Test management
- G06F11/3688—Test management for test execution, e.g. scheduling of test suites
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/30—Information retrieval; Database structures therefor; File system structures therefor of unstructured textual data
- G06F16/33—Querying
- G06F16/3331—Query processing
- G06F16/334—Query execution
- G06F16/3344—Query execution using natural language analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/20—Natural language analysis
- G06F40/279—Recognition of textual entities
- G06F40/289—Phrasal analysis, e.g. finite state techniques or chunking
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
- Y02D10/00—Energy efficient computing, e.g. low power processors, power management or thermal management
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Artificial Intelligence (AREA)
- General Health & Medical Sciences (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Data Mining & Analysis (AREA)
- Databases & Information Systems (AREA)
- Mathematical Physics (AREA)
- Computer Hardware Design (AREA)
- Quality & Reliability (AREA)
- Human Computer Interaction (AREA)
- Machine Translation (AREA)
Abstract
The application discloses a method, a device, equipment and a storage medium for determining reply sentences, which relate to the technical field of software testing and can be used for rapidly and accurately determining a positioning and checking result aiming at a test problem. The method comprises the following steps: word segmentation processing is carried out on the to-be-processed problem description statement and vector representation is carried out, so that a to-be-processed problem description vector corresponding to the to-be-processed problem description statement is obtained; determining target problem description sentences matched with the problem description sentences to be processed from the sample problem description sentences based on sample problem description vectors and the problem description vectors to be processed which correspond to the sample problem description sentences in the pre-trained software test language model; determining a target reply sentence corresponding to the target question description sentence from sample reply sentences corresponding to the sample question description sentences respectively; wherein the sample reply sentence includes a cause description sub-sentence and a measure description sub-sentence.
Description
Technical Field
The present invention relates to the field of software testing technologies, and in particular, to a method, an apparatus, a device, and a storage medium for determining reply sentences.
Background
At present, a tester often encounters some test problems in the process of testing software projects, and at this time, the tester needs to locate and check the test problems according to self experience and in combination with some auxiliary tools so as to determine the reasons for the test problems and propose measures for solving the test problems.
However, the above-described prior art is not high in accuracy of the determined localization check result and is not high in localization check efficiency for the localization check process of the test problem (including the process of determining the cause of the occurrence of the test problem and the process of determining the measure for solving the test problem) depending on the test experience of the tester.
Disclosure of Invention
The method, the device, the equipment and the storage medium for determining the reply sentence can rapidly and accurately determine the positioning and checking result aiming at the test problem.
In order to achieve the above purpose, the present application adopts the following technical scheme:
in a first aspect, the present application provides a method for determining a reply sentence, including: word segmentation processing is carried out on the to-be-processed problem description sentences, each to-be-processed problem description keyword in the to-be-processed problem description sentences is determined, and the to-be-processed problem description vectors corresponding to the to-be-processed problem description sentences are determined according to the text index vectors corresponding to each to-be-processed problem description keyword; determining target problem description sentences matched with the problem description sentences to be processed from the sample problem description sentences based on sample problem description vectors and the problem description vectors to be processed which correspond to the sample problem description sentences in the pre-trained software test language model; determining a target reply sentence corresponding to the target question description sentence from sample reply sentences corresponding to the sample question description sentences respectively; wherein the sample reply sentence includes a cause description sub-sentence and a measure description sub-sentence.
In the technical scheme provided by the application, various test problems possibly occurring in the software test process and positioning and checking results respectively corresponding to the various test problems can be arranged in advance, and various sample problem description sentences (one sample problem description sentence is used for describing one test problem) and sample reply sentences respectively corresponding to the various sample problem description sentences (one sample reply sentence comprises one reason description sub-sentence and one measure description sub-sentence, one reason description sub-sentence is used for describing reasons for the occurrence of the corresponding test problems, and one measure description sub-sentence is used for describing measures for solving the corresponding test problems) are summarized; and training in advance according to sample problem description vectors respectively corresponding to the sample problem description sentences to obtain a software test language model. When a tester encounters a certain test problem in the test process, the test problem can be described as a problem description sentence (namely a to-be-processed problem description sentence in the application) and a positioning and checking platform (corresponding to a determining device of a reply sentence in the application) is input, the positioning and checking platform can firstly perform word segmentation processing on the to-be-processed problem description sentence, determine each to-be-processed problem description keyword in the to-be-processed problem description sentence, and determine a to-be-processed problem description vector corresponding to the to-be-processed problem description sentence according to a text index vector corresponding to each to-be-processed problem description keyword; then, the positioning and checking platform can respectively match the to-be-processed problem description vector with sample problem description vectors respectively corresponding to each sample problem description sentence in the pre-trained software test language model, and determine the sample problem description sentence corresponding to the sample problem description vector with the highest matching degree as a target problem description sentence for the to-be-processed problem description sentence. Then, a sample reply sentence corresponding to the target question description sentence in each sample reply sentence can be determined to be a target reply sentence aiming at the to-be-processed question description sentence, and a reason description sub-sentence and a measure description sub-sentence in the target reply sentence can represent a positioning and checking result aiming at the test question corresponding to the to-be-processed question description sentence. It can be seen that, according to the method and the device, through the pre-summarized various sample question description sentences, various sample reply sentences and the pre-trained software test language model, the to-be-processed question description sentences can be automatically processed, the target reply sentences aiming at the to-be-processed question description sentences are determined, and the occurrence reasons and the solving measures of the test questions described by the to-be-processed question description sentences are described in the target reply sentences. Therefore, the method and the device can quickly and accurately determine the positioning and checking result aiming at the test problem.
In a second aspect, the present application provides a determination apparatus for reply sentence, including: a vector representation module, a question description sentence matching module, and a reply sentence matching module;
the vector representation module is used for carrying out word segmentation processing on the to-be-processed problem description sentences, determining each to-be-processed problem description keyword in the to-be-processed problem description sentences, and determining the to-be-processed problem description vector corresponding to the to-be-processed problem description sentences according to the text index vector corresponding to each to-be-processed problem description keyword;
the problem description sentence matching module is used for determining a target problem description sentence matched with the problem description sentence to be processed from the sample problem description sentences based on the sample problem description vector and the problem description vector to be processed which correspond to each sample problem description sentence in the pre-trained software test language model;
the reply sentence matching module is used for determining a target reply sentence corresponding to the target question description sentence from sample reply sentences corresponding to the sample question description sentences respectively; wherein the sample reply sentence includes a cause description sub-sentence and a measure description sub-sentence.
In a third aspect, the present application provides a determining device for reply sentences, including a memory, a processor, a bus, and a communication interface; the memory is used for storing computer execution instructions, and the processor is connected with the memory through a bus; when the determination device of the reply sentence is running, the processor executes computer-executable instructions stored in the memory to cause the determination device of the reply sentence to execute the determination method of the reply sentence as provided in the first aspect described above.
In a fourth aspect, the present application provides a computer-readable storage medium having instructions stored therein, which when executed by a computer, cause the computer to perform the method of determining a reply sentence as provided in the first aspect.
In a fifth aspect, the present application provides a computer program product comprising computer instructions which, when run on a computer, cause the computer to perform the method of determining a reply sentence as provided in the first aspect.
It should be noted that the above-mentioned computer instructions may be stored in whole or in part on a computer-readable storage medium. The computer readable storage medium may be packaged together with the processor of the determination device of the reply sentence or may be packaged separately from the processor of the determination device of the reply sentence, which is not limited in this application.
The description of the second, third, fourth and fifth aspects of the present application may refer to the detailed description of the first aspect; further, the advantageous effects described in the second aspect, the third aspect, the fourth aspect, and the fifth aspect may refer to the advantageous effect analysis of the first aspect, and are not described herein.
In the present application, the names of the above-mentioned devices or functional modules are not limited, and in actual implementation, these devices or functional modules may appear under other names. Insofar as the function of each device or function module is similar to the present application, it is within the scope of the present application and the equivalents thereof.
These and other aspects of the present application will be more readily apparent from the following description.
Drawings
Fig. 1 is a flow chart of a method for determining a reply sentence according to an embodiment of the present application;
FIG. 2 is a schematic diagram of an N-dimensional semantic vector space according to an embodiment of the present application;
FIG. 3 is a flowchart illustrating another method for determining reply sentences according to an embodiment of the present application;
fig. 4 is a schematic structural diagram of a determination device for reply sentences according to an embodiment of the present application;
fig. 5 is a schematic structural diagram of a determination device for reply sentences according to an embodiment of the present application.
Detailed Description
The following describes in detail a method, an apparatus, a device, and a storage medium for determining a reply sentence provided in an embodiment of the present application with reference to the accompanying drawings.
The term "and/or" is herein merely an association relationship describing an associated object, meaning that there may be three relationships, e.g., a and/or B, may represent: a exists alone, A and B exist together, and B exists alone.
The terms "first" and "second" and the like in the description and in the drawings are used for distinguishing between different objects or for distinguishing between different processes of the same object and not for describing a particular sequential order of objects.
Furthermore, references to the terms "comprising" and "having" and any variations thereof in the description of the present application are intended to cover a non-exclusive inclusion. For example, a process, method, system, article, or apparatus that comprises a list of steps or elements is not limited to only those listed but may optionally include other steps or elements not listed or inherent to such process, method, article, or apparatus.
It should be noted that, in the embodiments of the present application, words such as "exemplary" or "such as" are used to mean serving as an example, instance, or illustration. Any embodiment or design described herein as "exemplary" or "for example" should not be construed as preferred or advantageous over other embodiments or designs. Rather, the use of words such as "exemplary" or "such as" is intended to present related concepts in a concrete fashion.
In the description of the present application, unless otherwise indicated, the meaning of "a plurality" means two or more.
In addition, the technical scheme of the application is used for acquiring, storing, using, processing and the like data, and the data are in accordance with relevant regulations of national laws and regulations.
The positioning and checking process (including the process of determining the cause of the test problem and the process of determining the measure for solving the test problem) of the prior art depends on the test experience of the tester, so that the accuracy of the determined positioning and checking result is not high, and the positioning and checking efficiency is also not high.
In view of the problems in the prior art, the embodiment of the application provides a method for determining reply sentences, which can automatically process a description sentence of a problem to be processed, determine a target reply sentence for the description sentence of the problem to be processed, and describe occurrence reasons and solving measures of a test problem described for the description sentence of the problem to be processed in the target reply sentence. Therefore, the method and the device can quickly and accurately determine the positioning and checking result aiming at the test problem.
The method for determining the reply sentence provided by the embodiment of the application can be applied to a positioning and checking platform (particularly can be applied to a background server side of the positioning and checking platform), and the positioning and checking platform can process the to-be-processed problem description sentence uploaded by the tester to determine a positioning and checking result (namely a target reply sentence).
The method for determining the reply sentence provided by the embodiment of the application may be performed by the device for determining the reply sentence provided by the embodiment of the application, and the device may be implemented in a software and/or hardware manner and integrated in a device for determining the reply sentence for performing the method. The determination device of the reply sentence may be a server or a server cluster corresponding to the location investigation platform, for example.
The following describes a method for determining a reply sentence provided in the present application with reference to the drawings.
Referring to fig. 1, the method for determining a reply sentence provided in the embodiment of the present application includes S101-S103:
s101, word segmentation processing is carried out on the to-be-processed problem description sentences, each to-be-processed problem description keyword in the to-be-processed problem description sentences is determined, and the to-be-processed problem description vectors corresponding to the to-be-processed problem description sentences are determined according to the text index vectors corresponding to each to-be-processed problem description keyword.
The to-be-processed problem description statement may be a statement described by a tester according to a test problem encountered in a test process.
In one possible implementation manner, a tester can upload the description statement of the problem to be processed to a background server side of the positioning and checking platform through a client side provided by the positioning and checking platform, and then the background server side of the positioning and checking platform can call a word segmentation tool to perform word segmentation processing on the description statement of the problem to be processed, so as to determine each description keyword of the problem to be processed in the description statement of the problem to be processed. Illustratively, the problem description statement to be processed may be: "reason for TPS gradually decreasing when processing an insertion request", the keyword description of each problem to be processed can be obtained by word segmentation on the TPS by a word segmentation tool: "at", "processing", "inserting", "requesting", "time", "TPS", "gradually", "decreasing", "cause".
After determining each of the to-be-processed problem description keywords in the to-be-processed problem description sentence, vector representation can be performed on each of the to-be-processed problem description keywords. For example, embodiments of the present application may vector-represent each pending issue description keyword based on a Bag of words model (Bag-of-words model). The bag-of-words model may be a pre-trained vector representation model, which is used to represent the problem description keywords of different semantics as N vectors of different dimensions in an N-dimensional vector space (N is a positive integer, N is equal to the category of the problem description keywords). For the current to-be-processed problem description keywords in the to-be-processed problem description keywords, determining text index vectors corresponding to the current to-be-processed problem description keywords through a word bag model, and then carrying out vector addition operation on the text index vectors corresponding to the to-be-processed problem description keywords respectively to obtain to-be-processed problem description vectors corresponding to-be-processed problem description sentences. For example, the problem description keyword "insert" may be represented by a vector a, the problem description keyword "request" may be represented by a vector b, and if the problem description sentence to be processed is "insert request", the problem description vector to be processed is a+b.
Optionally, word segmentation processing is performed on the to-be-processed problem description sentence, and each to-be-processed problem description keyword in the to-be-processed problem description sentence is determined, including: invoking a pre-trained software test word segmentation model to segment the to-be-processed problem description sentence, and determining each text word in the to-be-processed problem description sentence; and screening the text word in the problem description statement to be processed based on a preset screening rule to obtain the problem description keywords to be processed.
The preset screening rule may be a predetermined screening rule. In one possible implementation, the preset screening rule may be to filter text word segments of parts of speech that are irrelevant to semantic analysis, such as, for example, terms of a co-word, an adverb, a conjunctive word, a preposition, etc. (e.g., text word segments such as "may be filtered"), and filter stop words in a pre-created stop vocabulary (e.g., stop vocabulary includes "causes").
By way of example, taking the example that the problem description sentence to be processed is "the cause of gradually decreasing TPS when processing an insertion request", each text word may be obtained by word segmentation of the problem description sentence by a pre-trained software test word segmentation model: "at", "processing", "inserting", "requesting", "time", "TPS", "gradually", "decreasing", "cause"; screening the text segmentation words based on a preset screening rule to obtain description keywords of each to-be-processed problem: "processing", "inserting", "requesting", "TPS", "tapering", "lowering".
In practical application, the test staff mostly describe the test problem in a spoken language, so that a great amount of redundant information exists in the word segmentation result obtained by the word segmentation tool. Based on this, the embodiment of the application can perform targeted filtering on the word segmentation result through the preset screening rule, so that vector representation is performed on each filtered to-be-processed problem description keyword, and the accuracy of subsequent semantic reasoning can be improved (namely, the accuracy of sample problem description vectors matched with the to-be-processed problem description vectors is improved, and the accuracy of target problem description sentences matched with the to-be-processed problem description sentences is improved).
Optionally, invoking a pre-trained software test word segmentation model to segment the to-be-processed problem description sentence, and before determining each text word segment in the to-be-processed problem description sentence, the method for determining the reply sentence further includes: acquiring segmentation words of each field in the software testing field; and updating the preset word segmentation model based on the word segmentation in each field, the word frequency corresponding to the word segmentation in each field and the word property corresponding to the word segmentation in each field, so as to obtain the software test word segmentation model.
The preset word segmentation model may be a pre-trained word segmentation model. In this embodiment of the present application, an existing word segmentation tool (for example, jieba word segmentation tool) may be directly used as the preset word segmentation model. Each domain word may be a domain specific noun extracted from a large number of texts of the software test domain (a domain text set), and a word frequency corresponding to each domain word may represent a frequency of occurrence of the domain word in the domain text set.
In the software testing field, there are some field proper nouns, which may be a combination of two or more text word segmentation, and if word segmentation processing is performed on the to-be-processed problem description sentence containing the field proper nouns through a preset word segmentation model, the field proper nouns are split into a plurality of text word segmentation, which is not beneficial to restoring the original semantics of the to-be-processed problem description sentence. Based on the above, in the embodiment of the application, on the basis of the preset word segmentation model, the preset word segmentation model is optimized through the word segmentation in each field to obtain the software test word segmentation model, so that the accurate segmentation of the to-be-processed problem description statement can be realized through the optimized software test word segmentation model. Specifically, a proper noun dictionary in the software testing field can be established based on the proper noun in each field, the standard format of the proper noun dictionary is 'field word-word frequency-part of speech', and then a preset word segmentation model is optimized based on each proper noun in the proper noun dictionary, so that the software testing word segmentation model is obtained.
For example, taking the problem description sentence to be processed as the "reason for gradually decreasing TPS when processing an insertion request" as an example, if the word segmentation processing is performed on the problem description sentence by a preset word segmentation model, the "insertion" and the "request" are divided into two text words, and the word segmentation is performed on the problem description sentence by a software test word segmentation model, so that the field word "insertion request" is divided into one independent text word.
S102, determining target problem description sentences matched with the problem description sentences to be processed from the sample problem description sentences based on sample problem description vectors and the problem description vectors to be processed which correspond to the sample problem description sentences in the pre-trained software test language model.
The software test language model may be a language model trained in advance according to each sample problem description sentence and a sample problem description vector corresponding to each sample problem description sentence. Each sample problem description sentence can be a large number of summarized problem description sentences (one sample problem description sentence is used for describing one test problem) which are formed by sorting various test problems possibly occurring in the software test process in advance. The sample problem description vector may be a multidimensional vector obtained by vector-representing the sample problem description statement.
After the positioning and checking platform determines the to-be-processed problem description vector, the to-be-processed problem description vector and sample problem description vectors respectively corresponding to the sample problem description sentences can be sequentially matched, and then the sample problem description sentence corresponding to the sample problem description vector with the highest matching degree of the to-be-processed problem description vector can be determined as the target problem description sentence for the to-be-processed problem description sentence. In one possible implementation manner, the embodiment of the application may match the to-be-processed problem description vector with each sample problem description vector through a pre-trained machine learning model, so as to determine a sample problem description vector with the highest matching degree with the to-be-processed problem description vector.
Optionally, determining the to-be-processed problem description vector corresponding to the to-be-processed problem description sentence according to the text index vector corresponding to each to-be-processed problem description keyword, including: for each to-be-processed problem description keyword, determining a current space dimension corresponding to the current to-be-processed problem description keyword from each space dimension of a semantic vector space, and determining a text index vector corresponding to the current space dimension as a text index vector corresponding to the current to-be-processed problem description keyword; and determining the description vector of the problem to be processed based on the text index vector corresponding to each description keyword of the problem to be processed and the word weight corresponding to each description keyword of the problem to be processed.
Wherein, different question description keywords correspond to different spatial dimensions, and different spatial dimensions correspond to different text index vectors.
In an exemplary embodiment of the present application, all text fragments with different semantics in each sample question description sentence may be rearranged in advance, and after a preprocessing step such as rescreening, all text fragments after preprocessing are represented as vectors with different spatial dimensions, so as to form a semantic vector space (for example, may be a semantic vector space shown in fig. 2). For each to-be-processed problem description keyword, the word weight corresponding to the current to-be-processed problem description keyword is used for representing the importance degree of the current to-be-processed problem description keyword in the whole to-be-processed problem description sentence, and the importance degree is mapped in a semantic vector space to represent the size of a text index vector corresponding to the current to-be-processed problem description keyword.
Optionally, determining the to-be-processed problem description vector based on the text index vector corresponding to each to-be-processed problem description keyword and the word weight corresponding to each to-be-processed problem description keyword includes: for each to-be-processed problem description keyword, determining word weight corresponding to the current to-be-processed problem description keyword based on word frequency of the current to-be-processed problem description keyword in each to-be-processed problem description keyword and inverse document word frequency of the current to-be-processed problem description keyword in a preset text set; performing de-duplication treatment on each to-be-treated problem description keyword to obtain each candidate problem description keyword; and weighting the text index vectors respectively corresponding to the candidate problem description keywords based on the word weights respectively corresponding to the candidate problem description keywords to obtain the problem description vectors to be processed.
The preset text set may be a predetermined text set, for example, a text set formed by each sample question description sentence.
In the embodiment of the application, the word weight corresponding to each to-be-processed problem description keyword can be determined based on a TF-IDF algorithm (a calculation algorithm based on statistics). Specifically, the word frequency TF of the current pending problem description keywords in each pending problem description keyword can be determined based on the TF algorithm in the TF-IDF algorithm i,j Determining the inverse document word frequency IDF of the current to-be-processed problem description keyword in the preset text set based on the IDF algorithm in the TF-IDF algorithm i,j 。
Exemplary, embodiments of the present application may determine the word frequency tf by expression (1) i,j Determining the inverse document word frequency idf by the expression (2) i,j Determining the word weight tf by expression (3) i,j *idf i,j ：
Wherein n is i,j Represents the frequency of occurrence of the ith pending problem description keyword in the pending problem description sentence j, Σ k n k,j Representing the number of all the pending problem description keywords in the pending problem description statement j, Σ k n k,j The normalization operation on word frequency is included, so that the influence caused by different lengths of different to-be-processed problem description sentences is avoided. |d| represents the total number of problem description keywords in the preset text set, |D i The I represents the number of the ith to-be-processed problem description keywords in the preset text set, and the denominator is added 1 in the expression (2) to ensure that no situation that the denominator is 0 occurs when new problem description keywords which are not in the preset text set occur by utilizing Laplacian smoothing.
For example, if each candidate problem description keyword of the to-be-processed problem description sentence includes "insert request", "TPS", "gradually", "decrease", and text index vectors corresponding to "insert request", "TPS", "gradually", "decrease" are vector a, vector b, vector c, and vector d, and word weights corresponding to each are A1, A2, A3, and A4, then the text index vectors corresponding to each candidate problem description keyword are weighted, and the obtained to-be-processed problem description vector is: a1+a2+b+a3+c+a4+d.
Alternatively, in another possible implementation, the present embodiments may also determine word weights based on a latent semantic analysis (Latent Semantic Analysis, LSA) model.
Optionally, before determining, from each sample question description sentence, a target question description sentence matched with the to-be-processed question description sentence based on a sample question description vector and a to-be-processed question description vector respectively corresponding to each sample question description sentence in the pre-trained software test language model, the method for determining the reply sentence provided in the embodiment of the present application further includes: acquiring various sample problem description sentences; for each sample problem description sentence, word segmentation processing is carried out on the current sample problem description sentence, each current sample problem description keyword in the current sample problem description sentence is determined, and according to the text index vector respectively corresponding to each current sample problem description keyword, the sample problem description vector corresponding to the current sample problem description sentence is determined; and generating a software testing language model according to each sample problem description statement and the sample problem description vectors corresponding to each sample problem description statement.
For the processing procedure of the sample problem description statement, reference may be made to the foregoing procedure of representing the problem description statement to be processed as the problem description vector to be processed, which is not described herein in detail.
Optionally, based on the sample problem description vector and the to-be-processed problem description vector respectively corresponding to each sample problem description sentence in the pre-trained software test language model, determining a target problem description sentence matched with the to-be-processed problem description sentence from each sample problem description sentence, including: based on cosine similarity between each sample problem description vector and the problem description vector to be processed, determining a target problem description vector matched with the problem description vector to be processed from each sample problem description vector, and determining a sample problem description sentence corresponding to the target problem description vector as a target problem description sentence.
Referring to fig. 2, an N-dimensional semantic vector space is provided in an embodiment of the present application. As shown in fig. 2, each vector dimension in the N-dimensional semantic vector space may correspond to a question description keyword (e.g., in fig. 2, the N-dimensional semantic vector space includes vector dimensions corresponding to N question description keywords, namely, word 1, word 2, … …, and … …). In this embodiment of the present application, for a current sample problem description statement in each sample problem description statement, a current sample problem description vector corresponding to the current sample problem description statement may be mapped to the N-dimensional semantic vector space. Specifically, the current sample problem description vector can be expressed as a multidimensional vector in an N-dimensional semantic vector space according to a text index vector corresponding to each sample problem description keyword in the current sample problem description vector and a word weight corresponding to each sample problem description keyword; similarly, each sample problem description statement may be represented as a vector in the N-dimensional semantic vector space. After determining the problem description vector to be processed, the problem description vector to be processed may be also represented as a vector in the N-dimensional semantic vector space, then cosine similarity between each sample problem description vector and the problem description vector to be processed (i.e., the cosine value of the included angle θ in fig. 2 is calculated), then, according to the cosine similarity, one sample problem description vector with the minimum cosine similarity (i.e., the minimum θ value) is determined as a target problem description vector matching the problem description vector to be processed, and a sample problem description sentence corresponding to the target problem description vector is determined as a target problem description sentence.
S103, determining a target reply sentence corresponding to the target question description sentence from sample reply sentences corresponding to the sample question description sentences respectively.
Wherein the sample reply sentence includes a cause description sub-sentence and a measure description sub-sentence. One reason description sub-sentence is used to describe the reason why the corresponding test problem occurs, and one measure description sub-sentence is used to describe the measure to solve the corresponding test problem.
For example, taking the example that the sample question description statement is "the reason that TPS gradually decreases when processing an insert request", the sample reply statement may be "when processing an insert class request, it is necessary to determine the index value of the primary key (for example, which table is inserted) before performing an insert operation on data to be processed, and the creation level of the index directory affects the query speed, thereby affecting the speed of processing the insert class request; the index directory is reconstructed. The method comprises the steps of determining an index value (such as which table is inserted) of a main key before insertion operation is carried out on data to be processed when an insertion type request is processed, wherein the creation level of an index directory influences the query speed, so that the speed of processing the insertion type request is influenced.
By integrating the above description, in the method for determining reply sentences provided in the embodiment of the present application, various test problems that may occur in a software test process and positioning and troubleshooting results corresponding to the various test problems respectively may be consolidated in advance, and various sample problem description sentences (one sample problem description sentence is used for describing one test problem) and sample reply sentences corresponding to the various sample problem description sentences (one sample reply sentence includes one cause description sub-sentence and one measure description sub-sentence, one cause description sub-sentence is used for describing a cause of the occurrence of the corresponding test problem, and one measure description sub-sentence is used for describing measures for solving the corresponding test problem) are summarized; and training in advance according to sample problem description vectors respectively corresponding to the sample problem description sentences to obtain a software test language model. When a tester encounters a certain test problem in the test process, the test problem can be described as a problem description sentence (namely a to-be-processed problem description sentence in the application) and a positioning and checking platform (corresponding to a determining device of a reply sentence in the application) is input, the positioning and checking platform can firstly perform word segmentation processing on the to-be-processed problem description sentence, determine each to-be-processed problem description keyword in the to-be-processed problem description sentence, and determine a to-be-processed problem description vector corresponding to the to-be-processed problem description sentence according to a text index vector corresponding to each to-be-processed problem description keyword; then, the positioning and checking platform can respectively match the to-be-processed problem description vector with sample problem description vectors respectively corresponding to each sample problem description sentence in the pre-trained software test language model, and determine the sample problem description sentence corresponding to the sample problem description vector with the highest matching degree as a target problem description sentence for the to-be-processed problem description sentence. Then, a sample reply sentence corresponding to the target question description sentence in each sample reply sentence can be determined to be a target reply sentence aiming at the to-be-processed question description sentence, and a reason description sub-sentence and a measure description sub-sentence in the target reply sentence can represent a positioning and checking result aiming at the test question corresponding to the to-be-processed question description sentence. It can be seen that, according to the method and the device, through the pre-summarized various sample question description sentences, various sample reply sentences and the pre-trained software test language model, the to-be-processed question description sentences can be automatically processed, the target reply sentences aiming at the to-be-processed question description sentences are determined, and the occurrence reasons and the solving measures of the test questions described by the to-be-processed question description sentences are described in the target reply sentences. Therefore, the method and the device can quickly and accurately determine the positioning and checking result aiming at the test problem.
Optionally, as shown in fig. 3, the embodiment of the present application further provides a method for determining a reply sentence, where the method includes the following steps:
s301, acquiring each sample question description sentence and a sample reply sentence corresponding to each sample question description sentence.
S302, for each sample question description sentence, word segmentation processing is carried out on the current sample question description sentence, each current sample question description keyword in the current sample question description sentence is determined, and according to the text index vector corresponding to each current sample question description keyword, the sample question description vector corresponding to the current sample question description sentence is determined.
S303, generating a software test language model according to each sample problem description statement and sample problem description vectors corresponding to each sample problem description statement.
S304, under the condition that a to-be-processed problem description sentence is received, calling a pre-trained software test word segmentation model to segment the to-be-processed problem description sentence, and determining each text word in the to-be-processed problem description sentence; and screening the text word in the problem description statement to be processed based on a preset screening rule to obtain the problem description keywords to be processed.
S305, for each to-be-processed problem description keyword, determining a current space dimension corresponding to the current to-be-processed problem description keyword from space dimensions of a semantic vector space, and determining a text index vector corresponding to the current space dimension as a text index vector corresponding to the current to-be-processed problem description keyword; and determining word weights corresponding to the current to-be-processed problem description keywords based on the word frequency of the current to-be-processed problem description keywords in each to-be-processed problem description keyword and the inverse document word frequency of the current to-be-processed problem description keywords in the preset text set.
S306, performing de-duplication treatment on the description keywords of each problem to be treated to obtain candidate description keywords of each problem; and weighting the text index vectors respectively corresponding to the candidate problem description keywords based on the word weights respectively corresponding to the candidate problem description keywords to obtain the problem description vectors to be processed.
S307, determining a target problem description vector matched with the problem description vector to be processed from the sample problem description vectors based on cosine similarity between the sample problem description vectors and the problem description vector to be processed through a pre-trained software test language model, and determining sample problem description sentences corresponding to the target problem description vector as target problem description sentences.
S308, determining a target reply sentence corresponding to the target question description sentence from sample reply sentences corresponding to the sample question description sentences respectively.
As shown in fig. 4, the embodiment of the present application further provides a device for determining a reply sentence, where the device may include: the vector representation module 11, the question-description-sentence matching module 21, and the answer-sentence matching module 31.
Wherein the vector representation module 11 performs S101 in the above-described method embodiment, the question-description-sentence matching module 21 performs S102 in the above-described method embodiment, and the answer-sentence matching module 31 performs S103 in the above-described method embodiment.
Specifically, the vector representation module 11 is configured to perform word segmentation processing on the to-be-processed problem description sentence, determine each to-be-processed problem description keyword in the to-be-processed problem description sentence, and determine a to-be-processed problem description vector corresponding to the to-be-processed problem description sentence according to a text index vector corresponding to each to-be-processed problem description keyword;
a question description sentence matching module 21, configured to determine, from among the sample question description sentences, a target question description sentence that matches the question description sentence to be processed, based on a sample question description vector and the question description vector to be processed, which correspond to each sample question description sentence in the pre-trained software test language model, respectively;
A reply sentence matching module 31, configured to determine a target reply sentence corresponding to the target question description sentence from sample reply sentences respectively corresponding to the sample question description sentences; wherein the sample reply sentence includes a cause description sub-sentence and a measure description sub-sentence.
Optionally, the vector representation module 11 is specifically configured to:
invoking a pre-trained software test word segmentation model to segment the to-be-processed problem description sentence, and determining each text word in the to-be-processed problem description sentence; and screening the text word in the problem description statement to be processed based on a preset screening rule to obtain the problem description keywords to be processed.
Optionally, the determining device of the reply sentence provided by the application further includes a training module; the training module is used for:
before the vector representation module 11 invokes a pre-trained software test word segmentation model to segment the to-be-processed problem description sentence and determines each text word in the to-be-processed problem description sentence, acquiring each field word in the software test field; and updating the preset word segmentation model based on the word segmentation in each field, the word frequency corresponding to the word segmentation in each field and the word property corresponding to the word segmentation in each field, so as to obtain the software test word segmentation model.
Optionally, the vector representation module 11 is specifically further configured to:
for each to-be-processed problem description keyword, determining a current space dimension corresponding to the current to-be-processed problem description keyword from each space dimension of a semantic vector space, and determining a text index vector corresponding to the current space dimension as a text index vector corresponding to the current to-be-processed problem description keyword; wherein, different question description keywords correspond to different space dimensions, and different space dimensions correspond to different text index vectors; and determining the description vector of the problem to be processed based on the text index vector corresponding to each description keyword of the problem to be processed and the word weight corresponding to each description keyword of the problem to be processed.
Optionally, the vector representation module 11 is specifically further configured to:
for each to-be-processed problem description keyword, determining word weight corresponding to the current to-be-processed problem description keyword based on word frequency of the current to-be-processed problem description keyword in each to-be-processed problem description keyword and inverse document word frequency of the current to-be-processed problem description keyword in a preset text set; performing de-duplication treatment on each to-be-treated problem description keyword to obtain each candidate problem description keyword; and weighting the text index vectors respectively corresponding to the candidate problem description keywords based on the word weights respectively corresponding to the candidate problem description keywords to obtain the problem description vectors to be processed.
Optionally, the question-description-sentence matching module 21 is specifically configured to: based on cosine similarity between each sample problem description vector and the problem description vector to be processed, determining a target problem description vector matched with the problem description vector to be processed from each sample problem description vector, and determining a sample problem description sentence corresponding to the target problem description vector as a target problem description sentence.
Optionally, the determining device of the reply sentence provided by the application further includes a training module; the training module is used for:
before the problem description sentence matching module 21 determines a target problem description sentence matched with the problem description sentence to be processed from among the sample problem description sentences based on the sample problem description vectors and the problem description vectors to be processed respectively corresponding to the sample problem description sentences in the pre-trained software test language model, acquiring each sample problem description sentence; and for each sample problem description sentence, word segmentation processing is carried out on the current sample problem description sentence, each current sample problem description keyword in the current sample problem description sentence is determined, and a sample problem description vector corresponding to the current sample problem description sentence is determined according to the text index vector corresponding to each current sample problem description keyword; in addition, a software test language model is generated according to each sample problem description sentence and the sample problem description vector corresponding to each sample problem description sentence.
Optionally, the determining means of the reply sentence may further include a storage module for storing a program code or the like of the determining means of the reply sentence.
As shown in fig. 5, the embodiment of the present application further provides a determination device of reply sentences, including a memory 41, a processor (such as 42-1 and 42-2 in fig. 5), a bus 43, and a communication interface 44; the memory 41 is used for storing computer-executed instructions, and the processor is connected with the memory 41 through the bus 43; when the determination device of the reply sentence is running, the processor executes the computer-executed instructions stored in the memory 41 to cause the determination device of the reply sentence to execute the determination method of the reply sentence as provided in the above-described embodiment.
In a particular implementation, the processor may include, as one embodiment, one or more central processing units (central processing unit, CPU), such as CPU0 and CPU1 shown in fig. 5. And as one example, the determination device of the reply sentence may include a plurality of processors, such as the processor 42-1 and the processor 42-2 shown in fig. 5. Each of these processors may be a single-Core Processor (CPU) or a multi-core processor (multi-CPU). A processor herein may refer to one or more devices, circuits, and/or processing cores for processing data (e.g., computer program instructions).
The memory 41 may be, but is not limited to, a read-only memory 41 (ROM) or other type of static storage device that can store static information and instructions, a random access memory (random access memory, RAM) or other type of dynamic storage device that can store information and instructions, or an electrically erasable programmable read-only memory (electrically erasable programmable read-only memory, EEPROM), a compact disc read-only memory (compact disc read-only memory) or other optical disc storage, optical disc storage (including compact disc, laser disc, optical disc, digital versatile disc, blu-ray disc, etc.), magnetic disk storage media or other magnetic storage devices, or any other medium that can be used to carry or store desired program code in the form of instructions or data structures and that can be accessed by a computer. The memory 41 may be stand alone and be connected to the processor via a bus 43. The memory 41 may also be integrated with the processor.
In a specific implementation, the memory 41 is used for storing data in the application and computer-executable instructions corresponding to executing a software program of the application. The processor may respond to various functions of the determination device of the sentence by running or executing a software program stored in the memory 41 and invoking data stored in the memory 41.
Communication interface 44, using any transceiver-like device, is used to communicate with other devices or communication networks, such as a control system, a radio access network (radio access network, RAN), a wireless local area network (wireless local area networks, WLAN), etc. The communication interface 44 may include a receiving unit to implement a receiving function and a transmitting unit to implement a transmitting function.
Bus 43 may be an industry standard architecture (industry standard architecture, ISA) bus, an external device interconnect (peripheral component interconnect, PCI) bus, or an extended industry standard architecture (extended industry standard architecture, EISA) bus, among others. The bus 43 may be classified into an address bus, a data bus, a control bus, and the like. For ease of illustration, only one thick line is shown in fig. 5, but not only one bus or one type of bus.
As an example, in connection with fig. 4, the vector representation module, the question-description-sentence matching module, and the answer-sentence matching module in the determination means of the answer sentence realize the same functions as those realized by the processor in fig. 5. When the determination means of the reply sentence includes a memory module, the function realized by the memory module is the same as that realized by the memory in fig. 5.
The explanation of the related content in this embodiment may refer to the above method embodiment, and will not be repeated here.
From the foregoing description of the embodiments, it will be apparent to those skilled in the art that, for convenience and brevity of description, only the above-described division of functional modules is illustrated, and in practical application, the above-described functional allocation may be implemented by different functional modules, that is, the internal structure of the device is divided into different functional modules, so as to implement all or part of the functions described above. The specific working processes of the above-described systems, devices and units may refer to the corresponding processes in the foregoing method embodiments, which are not described herein again.
The present embodiment also provides a computer-readable storage medium having instructions stored therein, which when executed by a computer, cause the computer to perform the method for determining a reply sentence provided by the above embodiment.
The computer readable storage medium may be, for example, but not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or a combination of any of the foregoing. More specific examples (a non-exhaustive list) of the computer-readable storage medium would include the following: an electrical connection having one or more wires, a portable computer diskette, a hard disk, a RAM, a ROM, an erasable programmable read-only memory (erasable programmable read only memory, EPROM), a register, a hard disk, an optical fiber, a CD-ROM, an optical storage device, a magnetic storage device, or any suitable combination of the foregoing, or any other form of computer readable storage medium known in the art. An exemplary storage medium is coupled to the processor such the processor can read information from, and write information to, the storage medium. In the alternative, the storage medium may be integral to the processor. The processor and the storage medium may reside in an application specific integrated circuit (application specific integrated circuit, ASIC). In the context of the present application, a computer-readable storage medium may be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.
The foregoing is merely a specific embodiment of the present application, but the protection scope of the present application is not limited thereto, and any changes or substitutions within the technical scope of the present disclosure should be covered in the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims.
Claims (10)
1. A method of determining a reply sentence, comprising:
word segmentation processing is carried out on the to-be-processed problem description sentences, each to-be-processed problem description keyword in the to-be-processed problem description sentences is determined, and the to-be-processed problem description vectors corresponding to the to-be-processed problem description sentences are determined according to the text index vectors corresponding to each to-be-processed problem description keyword;
determining target problem description sentences matched with the to-be-processed problem description sentences from the sample problem description sentences based on sample problem description vectors and the to-be-processed problem description vectors respectively corresponding to the sample problem description sentences in a pre-trained software test language model;
determining a target reply sentence corresponding to the target question description sentence from sample reply sentences respectively corresponding to the sample question description sentences; wherein the sample reply sentence includes a cause description sub-sentence and a measure description sub-sentence.
2. The method for determining reply sentences according to claim 1, wherein the word segmentation processing is performed on the question description sentences to be processed to determine each question description keyword to be processed in the question description sentences to be processed, comprising:
invoking a pre-trained software test word segmentation model to segment the to-be-processed problem description sentence, and determining each text word segmentation in the to-be-processed problem description sentence;
and screening out each text word in the to-be-processed problem description sentence based on a preset screening rule to obtain each to-be-processed problem description keyword.
3. The method for determining reply sentences according to claim 2, wherein said invoking a pre-trained software test word segmentation model performs word segmentation processing on said question-description sentences to be processed, said method further comprising, before determining each text word segmentation in said question-description sentences to be processed:
acquiring segmentation words of each field in the software testing field;
and updating a preset word segmentation model based on the word segmentation in each field, the word frequency corresponding to the word segmentation in each field and the word property corresponding to the word segmentation in each field, so as to obtain the software test word segmentation model.
4. The method for determining a reply sentence according to claim 1, wherein the determining a description vector of a to-be-processed question corresponding to the description sentence of the to-be-processed question according to the text index vectors respectively corresponding to the description keywords of the to-be-processed question includes:
for each to-be-processed problem description keyword, determining a current space dimension corresponding to the current to-be-processed problem description keyword from space dimensions of a semantic vector space, and determining a text index vector corresponding to the current space dimension as a text index vector corresponding to the current to-be-processed problem description keyword; wherein, different question description keywords correspond to different space dimensions, and different space dimensions correspond to different text index vectors;
and determining the description vector of the problem to be processed based on the text index vector corresponding to the description key words of the problem to be processed and the word weight corresponding to the description key words of the problem to be processed.
5. The method for determining a reply sentence according to claim 4, wherein the determining the question description vector to be processed based on the text index vector to which the question description keywords to be processed respectively correspond and the word weights to which the question description keywords to be processed respectively correspond includes:
For each to-be-processed problem description keyword, determining word weight corresponding to the current to-be-processed problem description keyword based on word frequency of the current to-be-processed problem description keyword in each to-be-processed problem description keyword and inverse document word frequency of the current to-be-processed problem description keyword in a preset text set;
performing de-duplication treatment on the description keywords of each problem to be treated to obtain candidate description keywords of each problem;
and weighting the text index vectors corresponding to the candidate problem description keywords respectively based on the word weights corresponding to the candidate problem description keywords respectively to obtain the problem description vectors to be processed.
6. The method according to claim 1, wherein the determining, based on the sample question description vectors and the to-be-processed question description vectors respectively corresponding to the respective sample question description sentences in the pre-trained software test language model, a target question description sentence matching the to-be-processed question description sentence from the respective sample question description sentences includes:
based on cosine similarity between each sample problem description vector and the problem description vector to be processed, determining a target problem description vector matched with the problem description vector to be processed from each sample problem description vector, and determining a sample problem description sentence corresponding to the target problem description vector as the target problem description sentence.
7. The method according to any one of claims 1 to 6, wherein before determining a target question description sentence matching with each of the to-be-processed question description sentences from among the sample question description sentences, based on the sample question description vectors and the to-be-processed question description vectors respectively corresponding to each of the sample question description sentences in the pre-trained software test language model, the method further comprises:
acquiring the description sentences of the various sample questions;
for each sample question description sentence, word segmentation processing is carried out on the current sample question description sentence, each current sample question description keyword in the current sample question description sentence is determined, and a sample question description vector corresponding to the current sample question description sentence is determined according to a text index vector corresponding to each current sample question description keyword;
and generating the software testing language model according to the sample question description sentences and the sample question description vectors respectively corresponding to the sample question description sentences.
8. A determination apparatus of a reply sentence, characterized by comprising:
the vector representation module is used for carrying out word segmentation processing on the to-be-processed problem description sentences, determining each to-be-processed problem description keyword in the to-be-processed problem description sentences, and determining the to-be-processed problem description vector corresponding to the to-be-processed problem description sentences according to the text index vector corresponding to each to-be-processed problem description keyword;
The problem description sentence matching module is used for determining a target problem description sentence matched with each to-be-processed problem description sentence from each sample problem description sentence based on a sample problem description vector and the to-be-processed problem description vector which correspond to each sample problem description sentence in a pre-trained software test language model;
the reply sentence matching module is used for determining a target reply sentence corresponding to the target question description sentence from sample reply sentences corresponding to the sample question description sentences respectively; wherein the sample reply sentence includes a cause description sub-sentence and a measure description sub-sentence.
9. A device for determining reply sentences, which is characterized by comprising a memory, a processor, a bus and a communication interface; the memory is used for storing computer execution instructions, and the processor is connected with the memory through the bus;
when the determination device of the reply sentence is running, the processor executes the computer-executable instructions stored in the memory to cause the determination device of the reply sentence to perform the determination method of the reply sentence as claimed in any one of claims 1 to 7.
10. A computer-readable storage medium having instructions stored therein, which when executed by a computer, cause the computer to perform the method of determining a reply sentence according to any one of claims 1-7.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311661659.7A CN117633187A (en) | 2023-12-05 | 2023-12-05 | Method, device, equipment and storage medium for determining reply sentence |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311661659.7A CN117633187A (en) | 2023-12-05 | 2023-12-05 | Method, device, equipment and storage medium for determining reply sentence |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN117633187A true CN117633187A (en) | 2024-03-01 |
Family
ID=90018019
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311661659.7A Pending CN117633187A (en) | 2023-12-05 | 2023-12-05 | Method, device, equipment and storage medium for determining reply sentence |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117633187A (en) |
-
2023
- 2023-12-05 CN CN202311661659.7A patent/CN117633187A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| WO2021189951A1 (en) | Text search method and apparatus, and computer device and storage medium | |
| CN108595619A (en) | A kind of answering method and equipment | |
| CN105069143B (en) | Extract the method and device of keyword in document | |
| CN110569335B (en) | Triple verification method and device based on artificial intelligence and storage medium | |
| CN110941951B (en) | Text similarity calculation method, text similarity calculation device, text similarity calculation medium and electronic equipment | |
| US20210042391A1 (en) | Generating summary content using supervised sentential extractive summarization | |
| CN110427626B (en) | Keyword extraction method and device | |
| CN110874528B (en) | Text similarity obtaining method and device | |
| CN111476026A (en) | Statement vector determination method and device, electronic equipment and storage medium | |
| CN110162752B (en) | Article judging and re-processing method and device and electronic equipment | |
| CN112183117A (en) | Translation evaluation method and device, storage medium and electronic equipment | |
| CN112395866B (en) | Customs clearance sheet data matching method and device | |
| CN115114937A (en) | Text acquisition method, device, computer equipment and storage medium | |
| CN114328894A (en) | Document processing method, document processing device, electronic equipment and medium | |
| CN118052233A (en) | Constraint method, device, medium and program product for large model response illusion | |
| CN115906817A (en) | Keyword matching method and device for cross-language environment and electronic equipment | |
| CN110287284B (en) | Semantic matching method, device and equipment | |
| CN115859964B (en) | Educational resource sharing method and system based on educational cloud platform | |
| CN118364117A (en) | Government affair question-answering method based on knowledge graph and related equipment | |
| CN111858899A (en) | Statement processing method, device, system and medium | |
| CN117633187A (en) | Method, device, equipment and storage medium for determining reply sentence | |
| CN117993489A (en) | Data set generation method, device, electronic equipment and computer readable medium | |
| CN116483971A (en) | User question intelligent answer method, financial system and storage medium | |
| CN112395865B (en) | Check method and device for customs clearance sheet | |
| CN110929501B (en) | Text analysis method and device |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |