基于决策树的回归测试子集选取

软件版本的频繁变更及测试资源的限制要求软件回归测试采用新的测试用例集合的生成和约简技术。为此，介绍基于决策树的回归测试子集的选取方法，将测试用例和测试需求作为一种知识表示系统，对测试知识表示系统进行约简，将约简后的系统构造成一棵决策树，由决策树获得被约简的回归测试子集。理论分析证明该方法复杂度较低。

Selection of Regression Test Subset Based on Decision Tree

Increasing number of software versions and the limitation of test resources challenge the traditional methods of regression test. This paper presents a new approach to select test subset for regression test by building the decision tree of test knowledge system. Test cases and test requirements are denoted as a test knowledge system in this method. The test knowledge system is reduced by eliminating redundant attributes. It builds the decision tree from the reduced test knowledge system. A test subset can be obtained from the decision tree. To implement the method by computer, it also gives the algorithm for reducing test knowledge system and the algorithm for building the decision tree.