Abstract

Nowadays, maintaining an oversized and evolving computer code involves longer and value for the project team. In software maintenance, a bug report is employed to seek out a fault location. Once a bug report is received, it is suiTable.to automatically denote out the files that developers should change to repair the bug. This work develops a brand-new information retrieval (IR) system that allows representing textual data by signals. This new way of implementing gives us the chance to use various mathematical tools from the signal theory like Wavelets Transforms, unused nowadays within the field of IR. This paper proposes Wavelets Transforms for Bug Localization (WTBugLoc), a mathematical approach of IR-based bug localization using wavelet techniques. The results of the conducted experiments on the SWT (Standard Widget Toolkit) Eclipse project confirm the effectiveness of the proposed approach. The experiments also show that WTBugLoc outperforms method using the Vector Space Model (VSM).