CN101814055B - Sampling method for test cases in clusters - Google Patents
Sampling method for test cases in clusters Download PDFInfo
- Publication number
- CN101814055B CN101814055B CN2010101398484A CN201010139848A CN101814055B CN 101814055 B CN101814055 B CN 101814055B CN 2010101398484 A CN2010101398484 A CN 2010101398484A CN 201010139848 A CN201010139848 A CN 201010139848A CN 101814055 B CN101814055 B CN 101814055B
- Authority
- CN
- China
- Prior art keywords
- test case
- statement
- test
- failure
- suspicious
- 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.)
- Expired - Fee Related
Links
- 238000012360 testing method Methods 0.000 title claims abstract description 228
- 238000005070 sampling Methods 0.000 title claims abstract description 66
- 238000000034 method Methods 0.000 title claims abstract description 34
- 238000012795 verification Methods 0.000 claims description 7
- 238000005516 engineering process Methods 0.000 abstract description 13
- 230000000694 effects Effects 0.000 description 16
- 230000003044 adaptive effect Effects 0.000 description 13
- 238000002474 experimental method Methods 0.000 description 8
- 238000013522 software testing Methods 0.000 description 6
- 238000001228 spectrum Methods 0.000 description 6
- 238000007689 inspection Methods 0.000 description 5
- 238000001914 filtration Methods 0.000 description 3
- 239000012467 final product Substances 0.000 description 2
- 238000010200 validation analysis Methods 0.000 description 2
- 230000004069 differentiation Effects 0.000 description 1
- 239000000203 mixture Substances 0.000 description 1
- 238000010187 selection method Methods 0.000 description 1
Images
Landscapes
- Debugging And Monitoring (AREA)
Abstract
The invention discloses a stamping method for test cases in clusters, which comprises the following steps of: selecting the test cases in the clusters to test based on a technology for filtrating the clusters in an observed software test technology; computing the reliability of excused sentences according to a test result; judging whether the excused sentences are doubtful sentences, wherein the doubtful sentences form a doubtful sentences set; computing the failure probability of remainder test cases in the clusters according to the doubtful sentences set; selecting the test case with the highest failure probability; testing the result and recomputing the reliability of each sentence; updating the doubtful sentences set; and repeating the step of sampling to the untested test cases till that the failure probability of all remainder test cases in the clusters are 0. As the results of the test cases after sampling do not need to be tested, the method simples the test cases, saves the time and the energy for manually testing the test cases, and guarantees that the failure test cases in the test cases to be sampled has significant proportion and the reliability value of the sentences can be used for helping test persons to perform error locating.
Description
Technical field
The invention belongs to the software test field; Relate to software test result's checking; At first test case is carried out cluster, from each type bunch, extract test case then and let the hand inspection test result, need hand inspection result's test use cases to practice thrift manpower through yojan according to execution route.The present invention is mainly used in the generation of test input and the execution of test case can be accomplished automatically, but test result needs to be a kind of sampling method for test cases in clusters under the scene of artificial comparison.
Background technology
Traditional method for testing software comprises three steps: 1) create the test input; 2) implementation of test cases; 3) checkout result.Software test needs the time and the manpower of labor, but it is an important means of guaranteeing software quality.In order to practice thrift testing cost, the researchist has designed some automatic tools and has helped generate test input, implementation of test cases and checkout result.Yet, in manipulation type test (like the beta test, the field test), the test result variation, description accurately in advance is clear, therefore needs manual work to go checking, thus the manpower of labor.
Can accomplish automatically to the generation of this test input and the execution of test case; But test result needs the situation of artificial checking; People such as David Leon have proposed the software testing technology (Observation-Based Testing) based on observation, comprise three steps: 1) test input well for all test case creation; 2) carry out all test cases, and note the execution route of each test case; 3) analyze the execution path information of each test case, select test case and let its test result of hand inspection whether meet demand.Therefore, be to need the test case number of hand inspection to practice thrift manpower based on the software testing technology of observing, but choose maximum failure use-cases simultaneously as far as possible to keep the validity of test through minimizing.
Calendar year 2001, humans such as William Dickinson have proved that through experiment the cluster filtering technique is a kind of effectively based on the software testing technology of observation, mainly comprises following three steps:
1. move all test cases, note execution path information.
2. the execution route according to test case comes cluster, and the test case that has similar execution route like this will be gathered in same type bunch.
From each type bunch the sampling, select test case has constituted the test use cases after the yojan.The tester only need verify these test cases and get final product.
Wherein the sampling method in the 3rd step can influence the effect of test use cases yojan significantly.Had the scholar proposed below several kinds of sampling methods:
1) one of random choose (One-per-cluster Sampling) from each type bunch: under optimal situation, the failure test case that the same program mistake causes is all gathered in same type bunch.Find out all program errors, only need get final product from test case of random choose each type bunch.But in fact the effect of cluster can not be so good, possibly mix the failure test case that successful test case and different programs mistake cause in class bunch.In addition, a failure test case that only causes according to this mistake, the programmer is difficult to carry out location of mistake.Therefore, the target of sampling is to find out maximum failure test cases as far as possible.
2) from each type bunch random choose a plurality of (N-per-cluster Sampling): this method is intended to find out more failure use-case through selecting more test case.But this method is the simple extension of " random choose is from each type bunch " just, and they all are simple random sampling in essence, do not utilize any information to come guide sampling.
3) adaptive sampling (Adaptive Sampling): at first from each type bunch test case of random choose, then the execution result of all these test cases is verified.If a test case has been failed, then all the residue test cases in its place type bunch all can be selected and verify; Otherwise, the success of this test case, then all residue test cases all are filtered in its place type bunch.People such as William Dickinson are more effective than above two kinds of simple random samplings through experiment showed, this self-adaptation selection method.Under the equirotal situation of yojan collection, this method can be picked out more failure test case.Yet the effect of this adaptive sampling technology depends on the test case that the first time, random choose went out to a great extent.If the test case that random choose goes out from a class bunch checking result is failure, so whole type bunch all can be selected, even successful test case has accounted for major part in such bunch; Equally, if the test case that random choose goes out from a class bunch checking result be successfully, so whole class bunch all can be abandoned, even the test case of failing in such bunch has accounted for major part.
Summary of the invention
Technical matters to be solved by this invention is: existing software testing technology based on observation is comprehensive inadequately in the test result inspection; Need to improve existing " adaptive sampling " technology; The execution spectrum information that makes full use of test case selected in each type bunch helps select test case, and when decision stops sampling from such bunch.
Technical scheme of the present invention is: a kind of sampling method for test cases in clusters after the test case of software test is carried out, obtains a plurality of types bunches according to the execution route cluster of test case; In class bunch, selecting test case verifies; According to the checking result of selected test case, calculate the confidence level of its performed statement, if the test case checking is passed through; The statement confidence level increases, otherwise then reduces; If a statement is with a low credibility in given threshold value, then this statement is suspicious statement, and suspicious statement constitutes suspicious statement set; Calculate the possibility of remaining each test case failure in such bunch according to suspicious statement set; It is high more that certain test case was carried out its failure possibility of more suspicious statement; Select that the highest test case of failure possibility, recomputate the confidence level of each statement then according to the checking result of this test case, upgrade suspicious statement set; And calculate the possibility of not verified residue test case failure, select the highest test case of next failure possibility and verify its result; Whole sampling process repeat like this; The failure possibility of all residue test cases all is 0 in such bunch; The performed statement that promptly remains test case is not all in suspicious statement set; The remaining test case of taking a sample is all no longer verified its result, the ratio of failure test case in the test case that about degeneracy of realization test case is guaranteed to be taken a sample.
Concrete steps of the present invention are following:
1) establishes and obtain m type of bunch C after the cluster
1, C
2... C
m, j representation class bunch numbering, initialization j=1;
2) confidence level of all statements of initialization is 0; From class bunch C
iIn select the maximum test case t of perform statement;
3) verification test cases: the execution result of tester's verification test cases t, obtain feedback information: this test case is failure, still success;
4) calculate the perform statement confidence level: according to the checking result of test case t; The confidence level of computing statement; The formula of computing statement confidence level is confidence (s)=passed (s)-failed (s), and promptly the confidence level of statement s equals to carry out s and the number of the test case passed through, deducts the test case number of carrying out s and failure; If t authentication failed just, then the confidence level of all statements of carrying out of t subtracts 1; Otherwise, t success, then the confidence level of all statements of carrying out of t adds 1;
5) statement identification: according to the confidence level of statement; Statement is identified as two kinds of suspicious statement and credible statements; Set a threshold value ConfidenceThreshold; Be abbreviated as CT, all confidence levels are identified as suspicious statement less than the statement of CT, and concrete formula is suspicious (S)={ s ∈ S|confidence (s)＜CT}; All confidence levels are identified as credible statement more than or equal to the statement of CT, and concrete formula is correct (S)={ s ∈ S|confidence (s)＞=CT}; Do not occur simultaneously
between suspicious statement collection and the credible statement collection
6) test case failure possibility tolerance: the statement set representations that test case t carried out is t (S)={ s ∈ S|t executes s}; The possibility of test case t failure is the number of its suspicious statement of carrying out, computing formula be failpossibility (t)=| t (S) ∩ suspicious (S) |;
7) the maximum test case of failure possibility in type of the choosing bunch residue test case:, choose the maximum test case t ' of possibility that wherein fails according to the failure possibility of the residue test case of calculating; If a plurality of test cases have maximum failure possibility, then adopt the method for random choose, select a test case t ';
8) if the failure possibility value of t ' is 0, then stop from class bunch C
jIn sampling, j=j+1, if j＜=m would return step 2), begin sampling from next type bunch; If j＞m, after promptly all types bunch all take a samples and to be finished, whole sampling process end; If the failure possibility value of t ' is then returned step 3) greater than 0, continue sampling from current type bunch.
The present invention is based on the cluster filtering technique in the software testing technology of observation, is a kind of test case cluster sampling technology.The present invention needs the test case number of artificial checking to practice thrift manpower through yojan, guarantees to pick out failure test case as much as possible simultaneously.During sampling verifying software test result; The present invention's remaining test case of taking a sample is all no longer verified its result; Realized the yojan of test case; Practiced thrift artificial validation test result's time and efforts, the failure test case accounts for significant proportion in the test case of guaranteeing again to be taken a sample simultaneously, and the confidence value of statement can be carried out location of mistake with helping the tester.
Description of drawings
Fig. 1 is a schematic flow sheet of the present invention.
Fig. 2 (a) is 3-per-cluster sampling, adaptive sampling and ESBS of the present invention experiment effect figure on experimental arrangement PrintTokens.
Fig. 2 (b) is 3-per-cluster sampling, adaptive sampling and ESBS of the present invention experiment effect figure on experimental arrangement PrintTokens2.
Fig. 2 (c) is 3-per-cluster sampling, adaptive sampling and ESBS of the present invention experiment effect figure on experimental arrangement Schedule.
Fig. 2 (d) is 3-per-cluster sampling, adaptive sampling and ESBS of the present invention experiment effect figure on experimental arrangement Schedule2.
Fig. 2 (e) is 3-per-cluster sampling, adaptive sampling and ESBS of the present invention experiment effect figure on experimental arrangement Replace.
Fig. 2 (f) is 3-per-cluster sampling, adaptive sampling and ESBS of the present invention experiment effect figure on experimental arrangement Space.
Embodiment
The present invention is based on the cluster filtering technique in the software testing technology of observation, in class bunch, select test case and verify, according to the checking result of selected test case; Calculate the confidence level of its performed statement; If the test case checking is passed through, statement is credible, otherwise then is suspicious; Calculate the possibility of remaining each test case failure in such bunch according to the statement confidence level; It is high more that certain test case was carried out its failure possibility of more suspicious statement; Select that the highest test case of failure possibility; Recomputate the confidence level of each statement then according to the checking result of this test case, and calculate the possibility of not verified residue test case failure, select next test case; Whole sampling process repeat like this; The failure possibility of all residue test cases all is 0 in such bunch; The performed statement that promptly remains test case is credible; Then the test case through the sampling checking is successfully that through checking result differentiation test case still is the failure test case, and the remaining test case of taking a sample is successfully test case, realizes the automatic sampling of such bunch test case.
Through or the statement of the test case of failure carry out information and be called as " execution frequency spectrum " in the location of mistake field; Because the present invention is based on by the execution spectrum information of verification test cases and selects the residue test case; So can be called based on the sampling method Execution Spectra Based Sampling that carries out frequency spectrum, be subsequently abbreviated as ESBS.
Like Fig. 1, step of the present invention is following:
1) establishes and obtain m type of bunch C after the cluster
1, C
2... C
m, j representation class bunch numbering, initialization j=1;
2) confidence level of all statements of initialization is 0; From class bunch C
jIn select the maximum test case t of perform statement
3) verification test cases: the execution result of tester's verification test cases t, obtain feedback information: this test case is failure, still success;
4) calculate the perform statement confidence level: according to the checking result of test case t; The confidence level of computing statement; The formula of computing statement confidence level is confidence (s)=passed (s)-failed (s), and promptly the confidence level of statement s equals to carry out s and the number of the test case passed through, deducts the test case number of carrying out s and failure; If t authentication failed just, then the confidence level of all statements of carrying out of t subtracts 1; Otherwise, t success, then the confidence level of all statements of carrying out of t adds 1;
5) statement identification: according to the confidence level of statement; Statement is identified as two kinds of suspicious statement and credible statements; Set a threshold value ConfidenceThreshold; Be abbreviated as CT, all confidence levels are identified as suspicious statement less than the statement of CT, and concrete formula is suspicious (S)={ s ∈ S|confidence (s)＜CT}; All confidence levels are identified as credible statement more than or equal to the statement of CT, and concrete formula is correct (S)={ s ∈ S|confidence (s)＞=CT}; Do not occur simultaneously
between suspicious statement collection and the credible statement collection
6) test case failure possibility tolerance: the statement set representations that test case t carried out is t (S)={ s ∈ S|t executes s}; The possibility of test case t failure is the number of its suspicious statement of carrying out, computing formula be failpossibility (t)=| t (S) ∩ suspicious (S) |;
7) the maximum test case of failure possibility in type of the choosing bunch residue test case:, choose the maximum test case t ' of possibility that wherein fails according to the failure possibility of the residue test case of calculating; If a plurality of test cases have maximum failure possibility, then adopt the method for random choose, select a test case t ';
8) if the failure possibility value of t ' is 0, then stop from class bunch C
jIn sampling, j=j+1, if j＜=m would return step 2), begin sampling from next type bunch; If j＞m, after promptly all types bunch all take a samples and to be finished, whole sampling process end.If the failure possibility value of t ' is then returned step 3) greater than 0, continue sampling from current type bunch.
All need not verified its result by the test case that sampling is come out, and have so just realized labor-saving purpose.Simultaneously, the confidence value of statement can be used for helper person and carries out location of mistake.
To combine concrete instance below, practical implementation of the present invention will be described.Supposing has 6 test cases in the class bunch; Statement that they were carried out and checking result in the future are as shown in table 1; Wherein why be " checking result in the future "; Be because the checking result of test case only picks out and verified just and know, but in the example for the number percent of the failure test case of picking out is described, in table 1, listed all test cases checking result in the future.We will therefrom select test case with the sampling method among the present invention and let the tester verify now.Target is: pick out the least possible test case, but comprise failure test case as much as possible.It is concrete that to select process as shown in table 2.
1) just begin, the confidence level of all statements is initialized as 0.Because threshold value CT=1, the confidence level of all statements is all less than CT, and promptly all statements all are suspicious.Therefore, t2 is picked out by first, because it has carried out maximum suspicious statements, the possibility value of its failure is 5.
Shown in that row that order in the table 2 is 1, t2 checking back is for passing through, so the statement s1 that carried out of t2, s2, and s4, s5, the confidence level of s6 all adds 1, and at this moment, the confidence level of having only s3 is less than CT, and s3 is unique suspicious statement.In remaining 5 test cases, have only t3 and t5 to carry out s3, so they have identical failure possibility value 1.
2) shown in that row that order in the table 2 is 2, test case of random choose supposes that t3 is selected from t3 and t5, because t3 is verified as failure, and the statement s2 that t3 carried out, the confidence level of s3 subtracts 1, and the suspicious statement of this moment is s2 and s3.At remaining 4 test case t1, t4, t5, among the t6, t4 has carried out s2, and t5 has carried out s3, so they have identical failure possibility 1.
3) shown in that row that order in the table 2 is 3, test case of random choose supposes that t4 is selected, because t4 is verified as failure from t4 and t5; The statement s2 that t4 carried out, the confidence level of s5 subtracts 1, and the suspicious statement of this moment is s2, s3; S5 is at remaining test case t1, t5; Among the t6, have only t5 to carry out suspicious statement s3, t1 and t6 do not carry out any suspicious statement.Therefore have only t5 to have failure possibility 1.
4) shown in that row that order in the table 2 is 4, t5 is selected, because t5 is verified as through, the statement s1 that t5 carried out; S3, the confidence level of s4 adds 1, and the suspicious statement of this moment is s2, s3; S5, remaining test case t1, t6 do not carry out any suspicious statement, and the possibility of promptly failing value all is 0.Such bunch sampling finishes.
Sum up: have 6 test cases such as tables 1 in such bunch, we have selected wherein four test case: t2, t3, t4; T5 sees table 2, lets the tester verify its result; (t1 t6) does not verify its result to two other test case, has practiced thrift tester's validation test result's time like this.(t3, t4), the test case of these 2 failures all has been selected and in table 2 to have the test case of 2 failures in the table 1.That is to say that test case only verifying under the situation of 4/6 test case, found the test case of 100% failure, both practiced thrift the time, do not influence the test effect again.
Statement that carried out six test cases of table 1 and they and checking result in the future
| Test case | The statement of carrying out | Checking result in the future |
| t1 | s1，s4，s6 | Through |
| t2 | s1，s2，s4，s5，s6 | Through |
| t3 | s2，s3 | Failure |
| t4 | s2，s5 | Failure |
| t5 | s1，s3，s4 | Through |
| t6 | s1，s6 | Through |
Table 2 is based on the process of specifically selecting of table 1 test case
Like Fig. 2 (a)-(d), we operate in existing two kinds of cluster samplings technology 3-per-cluster sampling (getting 3 samples in each bunch at random) and adaptive sampling (adaptive sampling) and ESBS of the present invention (technological based on the cluster sampling of carrying out frequency spectrum) on Siemens's experimental arrangement of 6 classics.Among each figure; The be selected test case number verified of horizontal ordinate representative accounts for the number percent (%tests selected) of original test case number, and the failure test case number that the ordinate representative is selected accounts for the number percent (%failures found) of original failure test case number.Under the identical situation of the next test case number that is selected, the failure test case number that finds is many more, explains that this sampling technique effect is good more.Promptly under the identical situation of abscissa value, the vertical tabular value of sitting is big more, and this sampling technique is good more.As can be seen from the figure; ESBS method of the present invention is at Printtokens; Printtokens, Schedule2, Replace; The effect of Space on 5 experimental arrangements all obviously is better than 3-per-cluster sampling and adaptive sampling, and only the effect on Schedule is not as good as adaptivesampling.Find after searching reason that the effect of ESBS of the present invention depends on the effect of location of mistake to a great extent, and the mistake in the Schedule program is not easy the location, thereby has influenced the effect that test case is selected.
Claims (2)
1. sampling method for test cases in clusters after the test case of software test is carried out, obtains a plurality of types bunches according to the execution route cluster of test case; It is characterized in that in class bunch, selecting the maximum test case of perform statement verifies; According to the checking result of selected test case, calculate the confidence level of its performed statement, if the test case checking is passed through; The statement confidence level increases; Otherwise then reduce, promptly the confidence level of statement s equals to carry out s and the number of the test case passed through, deducts the test case number of carrying out s and failure; If a statement is with a low credibility in given threshold value, then this statement is suspicious statement, and suspicious statement constitutes suspicious statement set; Calculate the possibility of remaining each test case failure in such bunch according to suspicious statement set; It is high more that certain test case was carried out its failure possibility of more suspicious statement; Choose the maximum test case t ' of possibility that wherein fails,, then adopt the method for random choose if a plurality of test case has maximum failure possibility; Recomputate the confidence level of each statement then according to the checking result of this test case; Upgrade suspicious statement set, and calculate the possibility of not verified residue test case failure, select the highest test case of next failure possibility and verify its result; Whole sampling process repeat like this; The failure possibility of all residue test cases all is 0 in such bunch; The performed statement that promptly remains test case is not all in suspicious statement set; The remaining test case of taking a sample is all no longer verified its result, the ratio of failure test case in the test case that about degeneracy of realization test case is guaranteed to be taken a sample.
2. a kind of sampling method for test cases in clusters according to claim 1 is characterized in that concrete steps are following:
1) establishes and obtain m type of bunch C after the cluster
1, C
2... C
m, j representation class bunch numbering, initialization j=1;
2) confidence level of all statements of initialization is 0; From class bunch C
jIn select the maximum test case t of perform statement;
3) verification test cases: the execution result of tester's verification test cases t, obtain feedback information: this test case is failure, still success;
4) calculate the perform statement confidence level: according to the checking result of test case t; The confidence level of computing statement; The formula of computing statement confidence level is confidence (s)=passed (s)-failed (s), and promptly the confidence level of statement s equals to carry out s and the number of the test case passed through, deducts the test case number of carrying out s and failure; If t authentication failed just, then the confidence level of all statements of carrying out of t subtracts 1; Otherwise, t success, then the confidence level of all statements of carrying out of t adds 1;
5) statement identification: according to the confidence level of statement; Statement is identified as two kinds of suspicious statement and credible statements; Set a threshold value ConfidenceThreshold; Be abbreviated as CT, all confidence levels are identified as suspicious statement less than the statement of CT, and concrete formula is suspicious (S)={ s ∈ S|confidence (s)＜CT}; All confidence levels are identified as credible statement more than or equal to the statement of CT, and concrete formula is correct (S)={ s ∈ S|confidence (s)＞=CT}; Do not occur simultaneously
between suspicious statement collection and the credible statement collection
6) test case failure possibility tolerance: the statement set representations that test case t carried out is t (S)={ s ∈ S|t executes s}; The possibility of test case t failure is the number of its suspicious statement of carrying out, computing formula be failpossibility (t)=| t (S) ∩ suspicious (S) |;
7) the maximum test case of failure possibility in type of the choosing bunch residue test case:, choose the maximum test case t ' of possibility that wherein fails according to the failure possibility of the residue test case of calculating; If a plurality of test cases have maximum failure possibility, then adopt the method for random choose, select a test case t ';
8) if the failure possibility value of t ' is 0, then stop from class bunch Cj, to take a sample, j=j+1, if j＜=m would return step 2), begin sampling from next type bunch; If j＞m, after promptly all types bunch all take a samples and to be finished, whole sampling process end; If the failure possibility value of t ' is then returned step 3) greater than 0, continue sampling from current type bunch.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010101398484A CN101814055B (en) | 2010-04-06 | 2010-04-06 | Sampling method for test cases in clusters |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2010101398484A CN101814055B (en) | 2010-04-06 | 2010-04-06 | Sampling method for test cases in clusters |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101814055A CN101814055A (en) | 2010-08-25 |
| CN101814055B true CN101814055B (en) | 2012-07-25 |
Family
ID=42621315
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2010101398484A Expired - Fee Related CN101814055B (en) | 2010-04-06 | 2010-04-06 | Sampling method for test cases in clusters |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101814055B (en) |
Families Citing this family (12)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN102736972B (en) * | 2011-03-30 | 2016-03-09 | 国际商业机器公司 | The test case treating Test Application carries out the method and system processed |
| CN102736013B (en) * | 2011-04-12 | 2015-08-05 | 安凯(广州)微电子技术有限公司 | A kind of idle condition method of testing of SoC chip, system and proving installation |
| CN102193864B (en) * | 2011-05-13 | 2014-02-26 | 南京大学 | Test case set optimization method of coverage-based error positioning technology |
| CN102253889B (en) * | 2011-08-07 | 2014-02-26 | 南京大学 | Method for dividing priorities of test cases in regression test based on distribution |
| CN102521131B (en) * | 2011-12-13 | 2014-10-29 | 南京大学 | Defect positioning test case reduction method based on dynamic basic block |
| CN102831055B (en) * | 2012-07-05 | 2015-04-29 | 陈振宇 | Test case selection method based on weighting attribute |
| CN103744789B (en) * | 2014-01-23 | 2017-01-11 | 中国科学院软件研究所 | Method of locating software errors by 3D surface representation |
| CN104050083B (en) * | 2014-06-17 | 2017-05-03 | 南京大学 | Condition/Decision-coverage-oriented test data automatic generation method |
| CN104090837B (en) * | 2014-06-19 | 2018-02-09 | 苏州市职业大学 | The method for generating test case |
| CN106776208B (en) * | 2016-12-02 | 2018-08-31 | 中国航天系统科学与工程研究院 | Fault Locating Method when a kind of running software |
| CN112905443A (en) * | 2019-12-04 | 2021-06-04 | 阿里巴巴集团控股有限公司 | Test case generation method, device and storage medium |
| CN114880242B (en) * | 2022-06-08 | 2023-01-13 | 北京百度网讯科技有限公司 | Test case extraction method, device, equipment and medium |
Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1746862A (en) * | 2004-09-09 | 2006-03-15 | 北京航空航天大学 | Software associated fault inspection |
| CN1746861A (en) * | 2004-09-09 | 2006-03-15 | 北京航空航天大学 | Dynamic software clustering test method |
-
2010
- 2010-04-06 CN CN2010101398484A patent/CN101814055B/en not_active Expired - Fee Related
Patent Citations (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN1746862A (en) * | 2004-09-09 | 2006-03-15 | 北京航空航天大学 | Software associated fault inspection |
| CN1746861A (en) * | 2004-09-09 | 2006-03-15 | 北京航空航天大学 | Dynamic software clustering test method |
Non-Patent Citations (2)
| Title |
|---|
| PAUL E. AMMANN AND JOHN C. KNIGHT.Data Diversity: An Approach to Software Fault Tolerance.《IEEE TRANSACTIONS ON COMPUTERS》.1988,第37卷(第4期),418-425. * |
| William Dickinson, David Leon, and Andy Podgurski.Finding Failures by Cluster Analysis of Execution Profiles.《Software Engineering, 2001.Proceedings of the 23rd International Conference on》.2001,339-348. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN101814055A (en) | 2010-08-25 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101814055B (en) | Sampling method for test cases in clusters | |
| CN102682760B (en) | Overlapped voice detection method and system | |
| CN106447046B (en) | A kind of Size Dwelling Design scheme evaluating method based on machine learning | |
| CN104133771B (en) | A kind of test method and device | |
| CN110362484A (en) | Localization method, device, electronic equipment and the storage medium of software multiple faults | |
| Chapellier et al. | The γ Doradus CoRoT target HD 49434-II. Frequency analysis of the CoRoT data | |
| CN103488562B (en) | Automated testing method and device | |
| CN112465124B (en) | Twin depth space-time neural network model acquisition/fault diagnosis method and device | |
| CN109634846B (en) | ETL software testing method and device | |
| CN102831055A (en) | Test case selection method based on weighting attribute | |
| CN114935703A (en) | Automatic testing method, device and system for frequency conversion assembly | |
| CN102445941A (en) | Method for automatically determining and analyzing interoperability test results of on-board equipment of train control system | |
| Iyama et al. | Automatically generating test scripts for gui testing | |
| Ye et al. | Adaptive board-level functional fault diagnosis using incremental decision trees | |
| CN110716820A (en) | Fault diagnosis method based on decision tree algorithm | |
| CN102521131A (en) | Defect positioning test case reduction method based on dynamic basic block | |
| CN107168883B (en) | API data automatic testing method | |
| CN109408772A (en) | To the restoration methods of the abnormal data in continuity data | |
| CN110287118B (en) | Test data generation method and software test method of test case | |
| CN102103539A (en) | Z-specification-based test case generating method | |
| CN104536880A (en) | GUI program testing case augmentation method based on symbolic execution | |
| CN116992362A (en) | Transformer fault characterization feature quantity screening method and device based on Xia Puli value | |
| Ramya et al. | Clustering based prioritization of test cases | |
| CN111552639B (en) | Software test comprehensive control method and system | |
| CN111078193A (en) | Software development method and system for data analysis system |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| C17 | Cessation of patent right | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20120725 Termination date: 20140406 |