Sampling method for test cases in clusters

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 field of software testing and relates to the verification of software testing results. Firstly, the test cases are clustered according to the execution path, and then the test cases are extracted from each cluster to allow manual inspection of the test results, and the tests that require manual inspection of the results are reduced. Use case sets to save manpower. The present invention is mainly applied to the scenario where the generation of test input and the execution of test cases can be completed automatically, but the test results need to be manually compared, and it is a clustering and sampling method for test cases. Background technique Traditional software testing methods include three steps: 1) creating test inputs; 2) executing test cases; 3) checking test results. Software testing needs a lot of time and manpower, but it is an important means to ensure software quality. In order to save testing costs, researchers hav...

AI Technical Summary

Problems solved by technology

The technical problem to be solved by the present invention is: the existing observation-based software testing technology is not comprehensive enough in the inspection of test results, and the existing "adaptive sampling" technology needs to be improved to make full use of the selected test results in each cluster. Use case execution spectrum information to help pick test cases and decide when to stop sampling from the class cluster

Images