Technological Evaluation and Software Bug Training using Genetic Algorithm and Time Convolution Neural Network (GA-TCN)

Abstract

One crucial step in addressing bugs is assigning a proper severity rating to each report. New bugs are being reported at such a high rate that the bug repository has grown significantly in size. The bug triaging procedure has become more subjective as the size of the bug repository has grown. Therefore, the bug triaging procedure requires a classification of the severity of a problem report. Several machine learning approaches to automatically rank bugs by severity have been proposed. These system performance falls short because insufficient feature patterns are extracted for classifier training. The proposed approach consists of preprocessing, normalization and training the model. For training the model the proposed approach uses GA-TCN.