基于遗传算法的软件质量预测  

Software Quality Prediction Model Based on Genetic Algorithm

作　　者：李旺彦[1] 李成楠[1] 

机构地区：[1]秦皇岛职业技术学院,河北秦皇岛066100

出　　处：《计算机仿真》2012年第3期385-388,共4页Computer Simulation

基　　金：学院科研项目(2010YA001);教育部高校硕士点基金(200801120007)

摘　　要：研究提高软件质量问题,软件质量是一种智力产品,质量度量属性较多,传统神经网络无法准确提取最优度量软件质量属性,预测准确率低。为了提高软件质量预测准确率,将遗传算法引入到软件质量度量属性选择中。首先采用遗传算法选择最优软件质量度量属性,然后将度量属性输入神经网络进行训练,建立软件质量预测模型。通过仿真对模型性能进行测试,结果表明,遗传神经网络对软件质量预测模型降低软件质量预测错误率,提高预测准确率,在理论和实际上都具有创新性。Software is a kind of intellectual products,and has strong complexity,invisible and uncertainty,these characteristics increase the prediction difficulty of software quality.In order to improve the prediction accuracy of software quality,the paper put forward a kind of genetic algorithm of software quality prediction model.First of all,the genetic algorithm with strong global search ability was used to measure the software quality attribute optimization and find the optimal attributes subset.Then,the optimal attributes subset was taken as the neural network＇s input,and through the neural network training,the software quality prediction model was built.Finally,the simulation experiment on the performance of the model test was carried out.The experimental results show that the genetic algorithm of software quality prediction model has smaller prediction error and higher predict accuracy.

关 键 词：软件质量 遗传算法 属性选择 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]