Building a Data Mining Based Software Reliability Estimation Model

Development of quality software is the ultimate goal of any software development organization. But one of the most challenging aspects of quality is that it can typically be measured only post-delivery. There have been many attempts to estimate software quality during development. Such estimates are definitely likely to aid in the engineering of quality software by providing useful insights to the project manager who can make informed decisions on resource allocation. The present research is an attempt to apply data mining technique of classification to aid in such reliability estimation. While quality has many facets, the research focuses on reliability. Most organizations are typically equipped with data pertaining to their past software releases and this data can be put to use in building models that can aid in predicting the reliability of software under development. The research uses data metrics data collected from SonarCloud to build a model that can aid in predicting the reliability of software. The model is built using the popular data mining tool weka.