Software reliability prediction via relevance vector regression

详细信息    查看全文

• 作者：Jungang Lou^a ; ^b ; ^{loujungang0210@hotmail.com" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; Yunliang Jiang^b ; ^{jylsy@hutc.zj.cn" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; Qing Shen^b ; ^{sq@hutc.zj.cn" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; Zhangguo Shen^b ; ^{szgxx@hutc.zj.cn" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; Zhen Wang^c ; ^{wangzhenqq@hotmail.com" class="auth_mail" title="E-mail the corresponding author}Author Vitae ; Ruiqin Wang^b ; ^{angelwrq@163.com" class="auth_mail" title="E-mail the corresponding author}Author Vitae
• 关键词：Software reliability model ; Relevance vector machine ; Mann&ndash ; Kendall test ; Paired T-test
• 刊名：Neurocomputing
• 出版年：2016
• 出版时间：19 April 2016
• 年：2016
• 卷：186
• 期：Complete
• 页码：66-73
• 全文大小：622 K

文摘

The aim of software reliability prediction is to estimate future occurrences of software failures to aid in maintenance and replacement. Relevance vector machines (RVMs) are kernel-based learning methods that have been successfully adopted for regression problems. However, they have not been widely explored for use in reliability applications. This study employs a RVM-based model for software reliability prediction so as to capture the inner correlation between software failure time data and the nearest m failure time data. We present a comparative analysis in order to evaluate the RVMs effectiveness in forecasting time-to-failure for software products. In addition, we use the Mann-Kendall test method to explore the trend of predictive accuracy as m varies. The reasonable value range of m is achieved through paired T-tests in 10 frequently used failure datasets from real software projects.