Software Bug Prediction using Machine Learning on JM1 Dataset

Ensuring the quality of software systems is essential for effective and efficient usage in complex software development procedures. A vital component of the whole procedure is the early identification and forecasting of potential defects or issues in software components. Using the NASA-curated JM1 dataset, this study examines the prediction accuracy of various machine learning techniques, such as Naïve Bayes, Decision Trees, Random Forest, Support Vector Machine, Logistic Regression, Artificial Neural Networks, and K-Nearest Neighbors for detecting defects in software. The research highlights the significance of early defect discovery in software development using machine learning. The experimental results showcase the Random Forest model as the most effective, with high accuracy (81%), precision, recall, and low Root-Mean-Square Error. Furthermore, this study includes a comparative analysis with prior works, offering valu-able insights into the performance and efficacy of the proposed methodology. The study concludes by suggesting future research directions.