基于机器学习的软件缺陷预测研究

Survey on Software Prediction Based on Machine Learning

基于机器学习的软件缺陷预测是一种有效的提高软件可靠性的方法。该方法基于软件模块的统计特性预测软件模块可能出现的缺陷数或是否容易出现缺陷。通过对软件模块缺陷状况的预测,软件开发组织可以将有限的资源集中于容易出现缺陷的模块,从而有效地提高软件产品的质量。基于机器学习的软件缺陷预测近年来出现了很多研究成果,文章概述该领域近年来的主要研究成果,并根据各方法的特点进行了分类。

Software Quality Prediction Based on Machine Learning is an effective way to improve software reliability. This technology predicts the number of faults in software module or whether a software module is fault-prone or not based on statistical feature of software module.According to the prediction of faults in software module,software developing organization can focus on fault-prone module to improve software quality effectively.This technology is now a proven way to improve software quality,improvement emerged in recent years about it.This paper will study these improvements by categories.