Trends in Software Engineering Processes using Deep Learning: A Systematic Literature Review
Álvaro Fernández Del Carpio, Leonardo Bermón Angarita
Abstract
In recent years, several researchers have applied machine learning techniques to several knowledge areas achieving acceptable results. Thus, a considerable number of deep learning models are focused on a wide range of software processes. This systematic review investigates the software processes supported by deep learning models, determining relevant results for the software community. This research identified that the most extensively investigated sub-processes are software testing and maintenance. In such sub-processes, deep learning models such as CNN, RNN, and LSTM are widely used to process bug reports, malware classification, libraries and commits recommendations generation. Some solutions are oriented to effort estimation, classify software requirements, identify GUI visual elements, identification of code authors, the similarity of source codes, predict and classify defects, and analyze bug reports in testing and maintenance processes.