Comparison of ML, Deep Learning and Bio-inspired Algorithms in Bug Triaging

Effective bug triaging is critical for efficient software development as bugs can cause software failures and unexpected behavior. The bug triaging process involves report summarization, severity classification, and assignment of bugs to developers. This paper introduced a novel developer recommendation algorithm designed to rank developers to reduce the bug tossing length. To automate and improve the accuracy of bug triaging, various machine learning models, deep learning models, and bio-inspired algorithms are utilized. In this project, we compared the effectiveness of bio-inspired algorithms with traditional machine learning algorithms such as SVM, Random Forest, Logistic Regression, K-Nearest Neighbors, Decision Tree, and Random Forest. We utilized open-source datasets such as the eclipse bug dataset and Firefox dataset for our analysis. We preprocessed the data and used the Artificial Bee Colony (ABC) algorithm for optimizing bug triaging. Our results indicate that the ABC algorithm yields more accurate results with fitness score 78.603 as compared to traditional machine learning algorithms in which highest accuracy obtained was 0.56 by decision tree.

Paper

Full text

PDF

Comparison of ML, Deep Learning and Bio-inspired Algorithms in Bug Triaging

Semantic Scholar · Computer Science · 2023

Abstract

Effective bug triaging is critical for efficient software development as bugs can cause software failures and unexpected behavior. The bug triaging process involves report summarization, severity classification, and assignment of bugs to developers. This paper introduced a novel developer recommendation algorithm designed to rank developers to reduce the bug tossing length. To automate and improve the accuracy of bug triaging, various machine learning models, deep learning models, and bio-inspired algorithms are utilized. In this project, we compared the effectiveness of bio-inspired algorithms with traditional machine learning algorithms such as SVM, Random Forest, Logistic Regression, K-Nearest Neighbors, Decision Tree, and Random Forest. We utilized open-source datasets such as the eclipse bug dataset and Firefox dataset for our analysis. We preprocessed the data and used the Artificial Bee Colony (ABC) algorithm for optimizing bug triaging. Our results indicate that the ABC algorithm yields more accurate results with fitness score 78.603 as compared to traditional machine learning algorithms in which highest accuracy obtained was 0.56 by decision tree.

Similar papers

© 2026 NYSGPT2525 LLC