基于测试需求的数据挖掘及测试数据生成  

Data Mining and Test Data Generation Based on Test Requirements

作　　者：赵一丁[1] 邵开丽[2] 李志民[1] 楚纪正[3] 

机构地区：[1]中原工学院计算机学院,郑州450007 [2]黄河科技学院信息工程学院,郑州450063 [3]北京化工大学信息科学与技术学院,北京100029

出　　处：《计算机测量与控制》2013年第8期2043-2045,共3页Computer Measurement &Control

基　　金：国家自然科学基金(71001082)资助

摘　　要：针对软件测试过程中的实践难点,提出了一种通过对用户实际业务的数据挖掘来生成测试数据的新方法;用户的实际业务数据包含了大量对于测试很重要的信息,测试用例设计之前挖掘大量的历史业务数据,进行多个步骤的数据处理得到"测试准备数据库",根据测试需求选取其中对应的数据作为测试数据的设计依据;对于成本较高的部分数据挖掘,测试数据生成与数据挖掘应该交互渐进,以减少数据挖掘的范围;通过实践检验,这种方法在多种情况下成功地解决了实践难题。In order to solve the problem of testing in practice, an innovative approach of test data generation by mining actual data from user was proposed. The user~ s actual data contains more important information for testing, a large number of actual history data should be obtained before designing test case. ＂Database for testing＂ can be built through multiple steps of data processing. According to the test re- quirements, corresponding data can be found from the database and become test data. As to the part of data mining with High cost, test data generation and data mining should be interactive progressive in order to reduce the scope of data mining. It has been proved by practice that this method is applicable to testing in more projects.

关 键 词：软件测试 数据挖掘 测试数据 过程工业 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]