Bug Summarization Severity Classification and Assignment for Automated Bug Resolution Process

Abstract

Bug resolution process is an important aspect of the software development life cycle(SDLC). The aim of bug resolution process (BRP) is to determine the bugs in software and fix it. These software bugs are introduced in software s during the different phases of SDLC process. The different strategies and mechanisms are considered during the evolution of software to overcome the propagation of bugs. A significant amount of time, cost and effort is put on the identification of bugs. It is observed that some bugs are not identified during the software evolution. These bugs can lead to failure of software s and unexpected behavior. So, prior to delivery of software s, every company ensure that the software is bug free and meets it expectation. Hence, to address and manage the bugs, bug tracking or reporting system are designed such as Mozilla, Eclipse etc. These bug tracking systems store the information related to bugs, called bug repositories. The valuable information regarding for fixing the bugs are described in repositories. This information can be used to automate the BRP and also help developers in terms of reduced time and effort. The BRP is described in terms of report summarization, severity classification and assignment. This thesis addresses the issues associated with bug report summarization, bug severity classification and bug assignment. Bug report summarization is the process of generating the short description of lengthy bugs. To resolve the bug, initially developer analyses and understands the content of bug report and make a summary set. This process is time consuming and tedious as large number of bugs are deposited in repositories per day. In turn, bug fixing time can be increased. Therefore, to automate and improve the accuracy rate of summarization task, a new summary subset selection technique is proposed to determine optimal summary subsets. This proposed technique is based particle swarm optimization (PSO) and ant colony optimization (ACO). The aim of proposed technique is to address data re