CN108197028A - A kind of regression test case sorting technique under the background towards Black-box Testing - Google Patents
A kind of regression test case sorting technique under the background towards Black-box Testing Download PDFInfo
- Publication number
- CN108197028A CN108197028A CN201810010235.7A CN201810010235A CN108197028A CN 108197028 A CN108197028 A CN 108197028A CN 201810010235 A CN201810010235 A CN 201810010235A CN 108197028 A CN108197028 A CN 108197028A
- Authority
- CN
- China
- Prior art keywords
- test case
- test
- classification
- text
- case
- 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.)
- Granted
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F11/00—Error detection; Error correction; Monitoring
- G06F11/36—Preventing errors by testing or debugging software
- G06F11/3668—Software testing
- G06F11/3672—Test management
- G06F11/3684—Test management for test design, e.g. generating new test cases
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Hardware Design (AREA)
- Quality & Reliability (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
- Debugging And Monitoring (AREA)
Abstract
The present invention provides the regression test case sorting techniques under a kind of background towards Black-box Testing, include the following steps:1) the corresponding text message of pretreatment test case;2) theme modeling is carried out to pretreated text using LDA (Latent Dirichlet Allocation) topic model technologies, is expressed as the fixed number of theme feature vector of text feature;3) it randomly selects partial test use-case and constructs Test oracles for it, and run on the software version after evolution, its classification is marked according to operation result;4) classification information based on theme feature vector and test case trains SVM classifier;5) using the corresponding theme feature vector of test case to be sorted as the input of grader, the classification of test case is exported.The present invention solves the software regression testing validation problem under the invisible situation of tested software source code, improves the degree of automation and Efficiency of Software Testing of software test.
Description
Technical field
It is surveyed the invention belongs to software testing technology field, and especially with regard to the recurrence under a kind of background towards Black-box Testing
Example sorting technique on probation.The method achieve tested software source code is invisible and Test oracles missing situation under software test
Automatic Verification enhances the feasibility of the regression test without Test oracles under Black-box Testing background, improves software recurrence
Testing efficiency.
Background technology
Software test is a key activities in software life-cycle, be mainly used for ensureing software quality with it is reliable
Property.Due to the upgrading of software, the reparation of software error and the variation of software runtime environment, software is in dynamic evolution.Software
It needs to carry out regression test to it once changing, to ensure the variation of software without introducing new mistake, and not to original
Some codes generate side effect.In addition, programmer establishes a kind of confidence also by regression test.As software life-cycle cost
One of activity of most expensive, regression test account for the 50% of software maintenance stage total cost, total to test the 80% of budget.In the limit
In programming and iteratively faster development mode, regression test consumes more test resources.The complexity of software and regression test
High test cost proposes new challenge to software test and maintenance personnel.Regression test resource consumption is effectively reduced,
Improving the efficiency of regression test has become soft project research field and industrial quarters urgent problem to be solved.
In view of the reasons such as the protection of intellectual property and trade secret, when software gives the third-party institution and is tested and assessed, source
Code is underground.Which increase the difficulty of its test.Existing survey is either run on the current software version of evolution
Whether the test case that example on probation or operation increase newly is required for examining the operation of tested software correct based on Test oracles.So
And the structure of Test oracles is completed by expertise, is taken time and effort very much.Such as:Verify medical image segmentation software
The professional knowledge for relying on doctor is just needed to build Test oracles, judges whether the medical image of segmentation is correct accordingly.
Original test use cases not necessarily meet the requirement of testing adequacy after Software Evolution, need to generate new test
Use-case, while to be also that each newly generated test case constructs Test oracles.Similarly, for the test in old version
The Test oracles of use-case construction can fail on software version after evolution, this just needs the software version based on evolution to be
Original test case reconfigures Test oracles.It is still to generate Test oracles by manual mode in current industrial quarters.This
Mean that software test needs to consume more test resources.With being continuously increased for software diversity and complexity, there is an urgent need for soft
The automatic Verification technology of part test, to meet industrial requirement.
Under regression test background, if do not construct Test oracles or only a small amount of test case construction Test oracles and
Can judge the classification of regression test case automatically will greatly reduce regression test energy consumption, improve regression test efficiency.Test is used
Example reflects test intention, and the error detection capability of similar test case is very close.This means that similar test
The classification of use-case is almost consistent.The corresponding text message of test case reflects test intention, collects the information and training
Disaggregated model judges newly-increased test case with original test case on the software version after evolution so as to fulfill automation
Classification.This for Test oracles lack and the unavailable situation of tested software source code under regression test provide it is a kind of efficiently, can
Capable solution.
As agile development and the continuous of iteratively faster development approach are popularized, the regression test of software becomes more frequently.
For this purpose, software regression testing needs to consume more test resources.In addition, Test oracles are difficult to construct and tested software source code
Invisibility both increase the difficulty of software regression testing.This so that traditional regression test case sorting technique is more next
The actual needs of software regression testing cannot more be met.
Invention content
Present invention aims at the regression test case sorting techniques provided under a kind of background towards Black-box Testing, solve quilt
Survey that software source code is invisible and Test oracles missing situation under software regression testing automatic Verification problem.It is used based on test
The characteristic information of example carries out mechanized classification to regression test case, realizes the automatic chemical examination of tested software under Black-box Testing background
Card so as to improve software regression testing efficiency, and then enhances the quality of software product.
The technical scheme is that:Towards the regression test case sorting technique under Black-box Testing background, history is surveyed
Example text message corresponding with newly-increased test case on probation is pre-processed, using LDA (Latent
DirichletAllocation) topic model technology carries out theme modeling to the text of pretreatment, further by test case pair
The text representation answered is the theme feature vector;The original test case in part and newly-increased test case are randomly selected, and for current
Software version constructs Test oracles for it, and the operation result based on the test case of selection on current software version marks its class
Not;Using the test case that marked as training set, Training Support Vector Machines (Support Vector Machine, SVM) point
Class model;Using the corresponding theme feature vector of test case to be sorted as the input of disaggregated model, its classification of final output.
To realize above-mentioned target, the present invention proposes the regression test case classification side under a kind of background towards Black-box Testing
Method, this method are as follows:
1) to original test case, text message corresponding with newly-increased test case pre-processes, and removes number, program
The keyword and specific character of language divide identifier title, by each word based on camel nomenclature (Camle-Case)
Its citation form is converted into, to reduce the vocabulary number that test case text includes;
2) theme modeling is carried out to pretreated text using LDA topic models technology, obtains each test case
Distribution of the corresponding text about theme, and the corresponding text representation of each test case is the theme feature vector;
3) 10% test case is respectively randomly choosed from original test case set and newly-increased test case set, for
Software version after evolution constructs Test oracles for the test case of selection, and is run on the software version after evolution selected
Whether test case unanimously marks the classification of test case according to operation result and its Test oracles, if unanimously, test case is
Correct use-case, on the contrary it is the use-case of mistake;
4) the corresponding theme feature vector of test case and classification information training svm classifier mould based on step 3) selection
Type;
5) the svm classifier model for selecting the corresponding theme feature vector of test case to be sorted as step 3) and generating
Input, the classification of final output test case to be sorted.
Further, wherein above-mentioned steps 1) it is as follows:
Step 1) -1:Initial state;
Step 1) -2:The corresponding text message of a test case is read by row;
Step 1) -3:Program language keyword in the style of writing sheet, operator are removed, word is divided based on camel nomenclature
It cuts, removal and the relevant stop-word of English language, and each word is converted into citation form;
Step 1) -4:Step 1) -2 and step 1) -3 are repeated, until having handled the corresponding text message of the test case, and
It saves it in disk file;
Step 1) -5:Repeat above-mentioned steps 1) -2, step 1) -3 and step 1) -4, until having pre-processed all history tests
Use-case and the corresponding text message of newly-increased test case;
Step 1) -6:The corresponding text message pretreatment of test case finishes;
Further, wherein above-mentioned steps 2) it is as follows:
Step 2) -1:Initial state;
Step 2) -2:The corresponding document of the pretreated each test case of step 1) is read in one by one, to the document into row vector
Change;
Step 2) -3:Repeat step 2) -2, until having handled all documents, most at last all document representations into document-spy
Levy word matrix;
Step 2) -4:The value of initial parameter α, β, k are inputted, they represent that the parameter of Dirichlet distributions, theme generation are single respectively
The probability of word and theme number carry out parameter Estimation, ultimately constructed LDA topic models using Gibbs sampling algorithms;
Step 2) -5:Document is corresponded to using LDA to each test case to handle, to obtain the theme probability point of document
Cloth represents that each element value in vector represents that the document is under the jurisdiction of the probability value of theme using vector;
Step 2) -6:Step 2) -5 are repeated, the theme probability distribution until obtaining all documents;
Step 2) -7:In number and its theme feature vector to disk file for preserving test case;
Step 2) -8:The theme feature vector construction of test case finishes;
Further, wherein above-mentioned steps 3) it is as follows:
Step 3) -1:The test case that 10% is respectively randomly choosed in original test case set and newly-increased test case set is made
Training set for train classification models;
Step 3) -2:For the new software version after evolution Test oracles are constructed for the test case that step 3) -1 selects;
Step 3) -3:Operating procedure 3 on new software version after evolution) -1 test case selected, and collect test and use
Example operation result information;
Step 3) -4:According to Test oracles and test result annotation step 3) -1 test case classification selected, if test knot
Fruit is consistent with Test oracles, which is correct use-case, and label is denoted as 1, and otherwise, for the use-case of mistake, label is denoted as-
1;
Step 3) -5:Step 3) -4 are repeated, the selectable test case until being labelled with;
Step 3) -6:Preserve the number and label information of test case;
Step 3) -7:The classification mark of test case finishes;
Further, above-mentioned steps 4) it is as follows:
Step 4) -1:Read step 3) selection test case number and label;
Step 4) -2:It is numbered by the test case of reading and obtains test case from the theme feature vector file that step 2) generates
Theme feature vector;
Step 4) -3:Repeat step 4) -1 and step 4) -2, until have read all test case theme features marked to
Amount and label information, using the theme feature vector of reading and its label as the input of train classification models;
Step 4) -4:Using radial basis function as kernel function, at the same the value of specified arrival rate σ, training svm classifier model;
Step 4) -5:Svm classifier Construction of A Model finishes;
Further, above-mentioned steps 5) it is as follows:
Step 5) -1:Select test case to be sorted, and the theme feature vector text generated by its number information from step 2)
The theme feature vector of the test case is obtained in part;
Step 5) -2:The input of svm classifier model generated using theme feature vector as step 4), according to the output valve of model
The classification of discriminating test use-case;
Step 5) -3:Test case kind judging to be sorted finishes.
The present invention is based on the regression test cases under the semantic information progress Black-box Testing background of test case to classify, substantially
Improve the automation efficiency of software regression testing;This method efficiently solves that tested software source code is invisible and test is pre-
Regression test validation problem under speech missing situation, this method are applicable not only on the software version after evolution to original survey
Example on probation is classified, while is also applied for increasing newly the classification of test case for the software version after developing;The present invention from
Test case corresponds to the pretreatment of text, is built to the theme feature vector based on topic model, then the training to disaggregated model,
It is carried out by the way of automation, greatly increases the automation efficiency of regression test, thus preferably control software matter
Amount.
Description of the drawings
Fig. 1 is the flow of the regression test case sorting technique under a kind of background towards Black-box Testing of the embodiment of the present invention
Figure.
Fig. 2 is the flow chart that test case corresponds to Text Pretreatment in Fig. 1.
Fig. 3 is the flow chart that test case corresponds to text subject modeling in Fig. 1.
Fig. 4 is the flow chart that test case classification marks in Fig. 1.
Fig. 5 is the flow chart of training svm classifier model in Fig. 1.
Fig. 6 is the flow chart of test case classification in Fig. 1.
Specific embodiment
For the clearer technology contents for understanding the present invention, spy lifts specific embodiment and coordinates institute's accompanying drawings explanation such as
Under.
Fig. 1 is the flow of the regression test case sorting technique under a kind of background towards Black-box Testing of the embodiment of the present invention
Figure.
Regression test case sorting technique under a kind of background towards Black-box Testing, which is characterized in that include the following steps:
S101 test cases correspond to Text Pretreatment, the text corresponding with newly-generated test case to original test case
Information is pre-processed, and removes digital, program language keyword and specific character, and based on camel nomenclature (Camle-
Case) divide identifier title, each word is converted into its citation form, to reduce the vocabulary that test case text includes
Number;
S103 test cases correspond to text subject modeling, and pretreated text is carried out using LDA topic models technology
Theme models, and obtains distribution of the corresponding text of each test case about theme, and each test case is corresponding
Text representation is the theme feature vector;
S105 test cases classification marks, original test case set and each random choosing in newly-increased test case set
10% test case is selected, and the test that each test case for selection is constructed on its software version after evolution is pre-
Speech, and selected test case is run on the software version after evolution.It is tested according to the operation result of test case with it pre-
Whether speech unanimously marks the classification of test case.If consistent, test case is correct use-case, otherwise is the use-case of mistake;
S107 trains svm classifier model, is trained with the corresponding theme feature vector of the test case of selection and classification information
Svm classifier model;
S109 test case classifications select the corresponding theme feature vector of test case to be sorted as svm classifier mould
The input of type exports the classification of test case to be sorted.
Fig. 2 is the flow chart that test case corresponds to Text Pretreatment.To original test case and newly-generated test case
Corresponding text message is pre-processed, and removes digital, program language keyword and specific character, and based on camel nomenclature
(Camle-Case) divide identifier title, each word is converted into its citation form, to reduce test case text packet
The vocabulary number contained.It is as follows:
Step 1:Initial state;Step 2:The corresponding text message of a test case is read by row;Step 3:Removing should
Program language keyword, operator in composing a piece of writing originally, are split word based on camel nomenclature, remove related to English language
Stop-word, and each word is converted into citation form;Step 4:Step 2 and step 3 are repeated, until having handled this
The corresponding text message of test case, and save it in disk file;Step 5:Repeat above-mentioned steps 2, step 3 and
Step 4, until having pre-processed all history test cases and the corresponding text message of newly-increased test case;Step 6:Test case
Corresponding text message pretreatment finishes.
Fig. 3 is the flow chart that test case corresponds to text subject modeling.Using LDA topic model technologies to pretreated
Text carries out theme modeling, obtains distribution of the corresponding text of each test case about theme, and each test is used
The corresponding text representation of example is the theme feature vector.It is as follows:
Step 1:Initial state;Step 2:The corresponding document of pretreated each test case is read in one by one, to this article
Shelves carry out vectorization;Step 3:Repeat step 2, until having handled all documents, most at last all document representations into document-
Feature Words matrix;Step 4:The value of initial parameter α, β, k are inputted, they represent the parameter of Dirichlet distributions, theme life respectively
Into the probability of word and theme number, parameter Estimation, ultimately constructed LDA topic models are carried out using Gibbs sampling algorithms;Step
5:Document is corresponded to using LDA to each test case to handle, to obtain the theme probability distribution of document, using vector table
Show, each element value in vector represents that the document is under the jurisdiction of the probability value of theme;Step 6:Step 5 is repeated, until obtaining
The theme probability distribution of all documents;Step 7:Preserve the number of test case and its theme feature vector to disk file;Step
Rapid 8:The theme feature vector construction of test case finishes.
Fig. 4 is the flow chart of test case classification mark.From original test case set and newly-increased test case set
The test case of each random selection 10% constructs Test oracles for the software version after evolution for the test case of selection, and
Selected test case is run on software version after evolution, whether unanimously mark is surveyed according to operation result and its Test oracles
The classification of example on probation if unanimously, test case is correct use-case, otherwise is the use-case of mistake.It is as follows:
Step 1:10% test is respectively randomly choosed in original test case set and newly-increased test case set
Training set of the use-case as train classification models;Step 2:It is the test that step 1 selects for the new software version after evolution
Use-case constructs Test oracles;Step 3:The test case that operating procedure 1 selects on new software version after evolution, and collect
Test case operation result information;Step 4:According to the test case classification that Test oracles and test result annotation step 1 select,
If test result is consistent with Test oracles, which is correct use-case, and label is denoted as 1, otherwise, the use for mistake
Example, label are denoted as -1;Step 5:Step 4 is repeated, the selectable test case until being labelled with;Step 6:It preserves and surveys
The number and label information of example on probation;Step 7:The classification mark of test case finishes.
Fig. 5 is the flow chart of training svm classifier model.With the corresponding Text eigenvector of the test case of selection and classification
Information trains svm classifier model.It is as follows:
Step 1:Read the number and label of the test case of selection;Step 2:It is numbered by the test case of reading from theme
The theme feature vector of the test case is obtained in feature vector file;Step 3:Step 1 and step 2 are repeated, until having read
All test case theme feature vectors and label information marked, and in this, as the input of train classification models;Step
4:Using radial basis function as kernel function, at the same the value of specified arrival rate σ, training svm classifier model;Step 5:Svm classifier model
Construction finishes.
Fig. 6 is the flow chart of test case classification.Select the corresponding theme feature vector of test case to be sorted as
The input of svm classifier model exports the classification of test case to be sorted.It is as follows:
Step 1:Test case to be sorted is selected, and the survey is obtained from theme feature vector file by its number information
The theme feature vector of example on probation;Step 2:Using theme feature vector as the input of disaggregated model, according to the output valve of model
The classification of discriminating test use-case;Step 3:Test case kind judging to be sorted finishes.
In conclusion the software regression testing verification that the present invention is solved under the invisible situation of tested software source code is asked
Topic, this method only need to construct Test oracles for a small amount of test case, using topic model and support vector machines technology to returning
Test case is classified, and the degree of automation and running efficiency of regression test not only greatly improved, but also is also reduced soft
Part regression test energy consumption, thus the preferably quality of control software product.
Claims (6)
1. the regression test case sorting technique under a kind of background towards Black-box Testing, which is characterized in that history test case
Text message corresponding with newly-increased test case is pre-processed, main using LDA (Latent Dirichlet Allocation)
It inscribes modelling technique and theme modeling is carried out to the text of pretreatment, the corresponding text representation of test case is further characterized number
Fixed theme feature vector, randomly selects the original test case in part and newly-increased test case, and for current software version
Test oracles are constructed for it, its classification is marked according to operation result of the test case of selection on current software version, with mark
The test case noted is as training set, training svm classifier model, by the corresponding theme feature vector of test case to be sorted
As the input of disaggregated model, its classification of final output;This method includes the following steps:
1) pretreatment of the corresponding text message of test case;
To original test case, text message corresponding with newly-increased test case pre-processes, and removes digital, program language
Keyword and specific character, and identifier title is divided based on camel nomenclature (Camle-Case), each word is converted
Into its citation form, to reduce the vocabulary number that test case text includes;
2) the theme modeling of test case;
Theme modeling is carried out to pretreated text using LDA topic models technology, it is corresponding to obtain each test case
Distribution of the text about theme, and the corresponding text representation of each test case is the theme feature vector;
3) the test case classification mark of training set;
10% test case is respectively randomly choosed from original test case set and newly-increased test case set, after evolution
Software version construct Test oracles for each test case of selection, and run on the software version after evolution selected
Whether test case unanimously marks the classification of test case according to operation result and its Test oracles, if unanimously, test case is
Correct use-case, on the contrary it is the use-case of mistake;
4) SVM model trainings;
The corresponding Text eigenvector of test case and classification information Training Support Vector Machines classification mould based on step 3) selection
Type;
5) test case classification;
The input of disaggregated model that the corresponding Text eigenvector of test case to be sorted is selected to be generated as step 3), finally
Export the classification of test case to be sorted.
2. the regression test case sorting technique under the background according to claim 1 towards Black-box Testing, which is characterized in that
In step 1), the text message of test case is pre-processed, to reduce the vocabulary number that test case text includes;It reads
The text message of original test case and newly-increased test case removes program language keyword, operator and and English in text
The relevant stop-word of language language, is split word based on camel nomenclature, each word is converted into citation form, preserves
All pretreated test case text messages.
3. the regression test case sorting technique under the background according to claim 1 towards Black-box Testing, which is characterized in that
In step 2), theme modeling carries out pretreated test case text message, and will test using topic model technology
Use-case text representation theme feature vector;It is as follows:Vectorization is carried out to pretreated test case text, by it
It is expressed as text-Feature Words matrix, carries out parameter Estimation using Gibbs sampling algorithms, construct LDA topic models, utilize the mould
Type obtains the theme probability distribution of document, and most the text message of test case is expressed as theme feature vector at last.
4. the regression test case sorting technique under the background according to claim 1 towards Black-box Testing, which is characterized in that
In step 3), the classification of test case is marked, using the test case marked as training set, standard is done for train classification models
It is standby;It is as follows:10% survey is respectively randomly choosed from original test case set and newly-increased test case set
Example on probation constructs Test oracles, and run the test case of selection for test case of the software version after evolution for selection,
The classification of test case whether is unanimously marked based on operation result and Test oracles.
5. the regression test case sorting technique under the background according to claim 1 towards Black-box Testing, which is characterized in that
In step 4), the disaggregated model of training test case;It is as follows:Believed with the label of the test case of step 3) selection
The input of breath and its theme feature vector as train classification models, using radial basis function as kernel function, while specified arrival rate σ
Value, training svm classifier model.
6. the regression test case sorting technique under the background according to claim 1 towards Black-box Testing, which is characterized in that
In step 5), classified using the svm classifier model in step 4) to test case to be sorted;It is as follows:
For test case to be sorted, its corresponding theme feature vector is read from step 2), is produced the vector as step 4)
The input of raw svm classifier model, the classification information of final output test case to be sorted.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810010235.7A CN108197028B (en) | 2018-01-05 | 2018-01-05 | Regression test case classification method for black box test background |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201810010235.7A CN108197028B (en) | 2018-01-05 | 2018-01-05 | Regression test case classification method for black box test background |
Publications (2)
Publication Number | Publication Date |
---|---|
CN108197028A true CN108197028A (en) | 2018-06-22 |
CN108197028B CN108197028B (en) | 2020-12-01 |
Family
ID=62588370
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201810010235.7A Expired - Fee Related CN108197028B (en) | 2018-01-05 | 2018-01-05 | Regression test case classification method for black box test background |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN108197028B (en) |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110297764A (en) * | 2019-05-30 | 2019-10-01 | 北京百度网讯科技有限公司 | Loophole test model training method and device |
CN111951229A (en) * | 2020-07-22 | 2020-11-17 | 国网安徽省电力有限公司电力科学研究院 | Small hardware fitting image data set training method based on full connection layer augmentation |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2001034466A (en) * | 1999-07-16 | 2001-02-09 | Nec Corp | Method for developing program |
CN101908020A (en) * | 2010-08-27 | 2010-12-08 | 南京大学 | Method for prioritizing test cases based on classified excavation and version change |
CN105912716A (en) * | 2016-04-29 | 2016-08-31 | 国家计算机网络与信息安全管理中心 | Short text classification method and apparatus |
CN107102939A (en) * | 2016-11-09 | 2017-08-29 | 中国矿业大学 | A kind of regression test case automatic classification method |
-
2018
- 2018-01-05 CN CN201810010235.7A patent/CN108197028B/en not_active Expired - Fee Related
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
JP2001034466A (en) * | 1999-07-16 | 2001-02-09 | Nec Corp | Method for developing program |
CN101908020A (en) * | 2010-08-27 | 2010-12-08 | 南京大学 | Method for prioritizing test cases based on classified excavation and version change |
CN105912716A (en) * | 2016-04-29 | 2016-08-31 | 国家计算机网络与信息安全管理中心 | Short text classification method and apparatus |
CN107102939A (en) * | 2016-11-09 | 2017-08-29 | 中国矿业大学 | A kind of regression test case automatic classification method |
Non-Patent Citations (1)
Title |
---|
谢晓园等: "演化测试技术的研究 ", 《计算机科学与探索》 * |
Cited By (2)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110297764A (en) * | 2019-05-30 | 2019-10-01 | 北京百度网讯科技有限公司 | Loophole test model training method and device |
CN111951229A (en) * | 2020-07-22 | 2020-11-17 | 国网安徽省电力有限公司电力科学研究院 | Small hardware fitting image data set training method based on full connection layer augmentation |
Also Published As
Publication number | Publication date |
---|---|
CN108197028B (en) | 2020-12-01 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110209764B (en) | Corpus annotation set generation method and device, electronic equipment and storage medium | |
CN107239529B (en) | Public opinion hotspot category classification method based on deep learning | |
KR20200014510A (en) | Method for providing prediction service based on mahcine-learning and apparatus thereof | |
CN104285224B (en) | Method for classifying to text | |
US11551171B2 (en) | Utilizing natural language processing and machine learning to automatically generate proposed workflows | |
CN110069627A (en) | Classification method, device, electronic equipment and the storage medium of short text | |
CN112070138A (en) | Multi-label mixed classification model construction method, news classification method and system | |
Bhartiya et al. | Artificial neural network based university chatbot system | |
Singh et al. | Rule-based system for automated classification of non-functional requirements from requirement specifications | |
JP2020024534A (en) | Image classifier and program | |
CN106445908A (en) | Text identification method and apparatus | |
CN109284374A (en) | For determining the method, apparatus, equipment and computer readable storage medium of entity class | |
Sharp et al. | Toward Semi-autonomous Information: Extraction for Unstructured Maintenance Data in Root Cause Analysis | |
Varma et al. | Flipper: A systematic approach to debugging training sets | |
CN111506504A (en) | Software development process measurement-based software security defect prediction method and device | |
CN108197028A (en) | A kind of regression test case sorting technique under the background towards Black-box Testing | |
CN112035345A (en) | Mixed depth defect prediction method based on code segment analysis | |
US11010692B1 (en) | Systems and methods for automatic extraction of classification training data | |
CN104536880B (en) | Gui program test case amplification method based on semiology analysis | |
CN111126038B (en) | Information acquisition model generation method and device and information acquisition method and device | |
Kobylinski | PoliTa: A multitagger for Polish. | |
CN117313736A (en) | Implicit demand recognition method and device for comment text | |
CN109657710B (en) | Data screening method and device, server and storage medium | |
KR102406961B1 (en) | A method of learning data characteristics and method of identifying fake information through self-supervised learning | |
CN113901793A (en) | Event extraction method and device combining RPA and AI |
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 | ||
GR01 | Patent grant | ||
GR01 | Patent grant | ||
CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20201201 Termination date: 20220105 |
|
CF01 | Termination of patent right due to non-payment of annual fee |