中文文本纠错软件测试用例的最小化方法

Minimization Method of Test Case for Chinese Text Error Correction Software

针对目前尚未存在有效的中文文本纠错软件测试用例最小化方法的情况,设计了两种中文文本纠错软件测试用例集最小化方法:基于多目标遗传算法的中文文本纠错软件测试用例集最小化方法(Minimization Method based on Multi-objective Genetic Algorithm of Test case for Chinese text error Correction Software,MMG-CCS)和基于TextRank模型的中文文本纠错软件测试用例最小化方法(Minimization Method based on TextRank of Test case for Chinese text error Correction Software,MMT-CCS).MMG-CCS基于问题特点设计了一种中文文本纠错软件测试用例集覆盖度模型,并将测试用例集大小及覆盖度作为目标函数,为了高效求解该问题,MMG-CCS同时对遗传算子进行了改进.此外,MMT-CCS方法能将测试用例集最小化问题映射为图解问题,并根据问题特点对TextRank模型进行了针对性改进.实验结果表明,MMG-CCS和MMT-CCS对测试用例集的缩减程度高,且在不同中文文本纠错软件上对于不同中文文本纠错软件测试用例集均能保持相同的测试效果.

To address the lack of effective test case minimization methods for Chinese text error correction software,two test case set minimization methods for Chinese text error correction software are designed:Minimization Method based on Multi-objective Genetic Algorithm of Test case for Chinese text error Correction Software(MMG-CCS)and the Minimization Method based on TextRank model for Chinese text error Correction Software(MMG-CCS).MMG-CCS designs a Chinese text error correction software test case set coverage model based on the characteristics of the problem,and takes the test case set size and coverage as the objective function,in order to solve the problem efficiently,MMG-CCS also improves the genetic operator.In addition,the MMT-CCS method can map the test case set minimization problem to a graphical problem and improve the TextRank model according to the problem characteristics.The experimental results show that MMG-CCS and MMT-CCS can efficiently reduce the test case set and maintain the same test results for different Chinese text error correction software test case sets on different Chinese text error correction software.