Analysis and Prediction of Open Bugs Using Machine Learning Algorithms
- May 24, 2018
- Posted by: RSIS
- Category: Computer Science and Engineering
International Journal of Research and Scientific Innovation (IJRSI) | Volume V, Issue V, May 2018 | ISSN 2321–2705
Analysis and Prediction of Open Bugs Using Machine Learning Algorithms
Sachin A S, Dr. Rajashree Shettar
Department of Computer Science and Engineering, R V College of Engineering, Mysuru Road, Bengaluru, Karnataka, India.
Abstract– There are many fault tracking repositories, some of them are YouTrack, Bugzilla, MantisBT and Atlassian JIRA. Atlassian JIRA repository has been used in this study, as it is extensively accepted by most of the software companies. This repository contains significant information of many projects. Each project has different kinds of issues such as bug(faults) reports, enhancement required to an existing feature, and new feature of the product and task that needs to be done. This paper focuses on analysing the previously raised bug report(history) to understand the correlation and dependability of the attributes like number of bugs created per day, their priority, number of days or hours taken to resolve etc., The data is then processed into a new format which will comply to machine learning algorithms. Different machine learning approaches such as linear regression, support vector machine, K-nearest neighbour, and generalized linear models (Ridge, Lasso and ElasticNet) will be applied on to the data to evaluate and identify the effective and efficient machine learning algorithm for prediction of possible open bugs per day. The best learning algorithm will then be used to predict open bugs per day and make better estimation of time and resource allocation to resolve the issues.
Index Terms –Open bug prediction, JIRA, Machine learning, Project management, Software Quality.
I. INTRODUCTION
These days, for accelerating software maintenance and advancement most of the companies are using the fault or bug tracking systems to organize and keep track of bugs related to software projects. Such systems are precious for developers, testers and different stakeholders, as they use the system to comfortably report different kind of faults. Faults could be of different categories such as internal tester identified defects/faults, customer identified faults, faults identified by release testing team and so on. A fault is a flaw in the software which shows an unanticipated reaction or behavior of software system [1].
