Authors:
Rafael Meneses Santos
1
;
Israel Meneses Santos
2
;
Methanias Colaço Rodrigues Júnior
2
and
Manoel Gomes de Mendonça Neto
1
Affiliations:
1
Graduate Program in Computer Science, Federal University of Bahia, Salvador, Brazil
;
2
Department of Information Systems, Federal University of Sergipe, Itabaiana, Brazil
Keyword(s):
Mining Software Repositories, Self-admitted Technical Debt, Long Short-term Memory, Neural Networks, Deep Learning, Word Embedding.
Abstract:
Context: In software development, new functionalities and bug fixes are required to ensure a better user experience and to preserve software value for a longer period. Sometimes developers need to implement quick changes to meet deadlines rather than a better solution that would take longer. These easy choices, known as Technical Debt, can cause long-term negative impacts because they can bring extra effort to the team in the future. Technical debts must be managed and detected so that the team can evaluate the best way to deal with them and avoid more serious problems. One way to detect technical debts is through source code comments. Developers often insert comments in which they admit that there is a need to improve that part of the code later. This is known as Self-Admitted Technical Debt (SATD). Objective: Evaluate a Long short-term memory (LSTM) neural network model combined with Word2vec for word embedding to identify design and requirement SATDs from comments in source code.
Method: We performed a controlled experiment to evaluate the quality of the model compared with two language models from literature and LSTM without word embedding in a labelled dataset. Results: The results showed that the LSTM model with Word2vec have improved in recall and f-measure. The LSTM model without word embedding achieves greater recall, but perform worse in precision and f-measure. Conclusion: Overall, we found that the LSTM model and word2vec can outperform other models.
(More)