A Dual-Network MLSTM Model for Software Defect Density Prediction

Defect density prediction is essential for maintaining software quality in software engineering. This study seems to address a significant research gap by offering an in-depth evaluation of Machine Learning (ML) and Deep Learning (DL) approaches for predicting defect density. We used seventeen datasets that were taken from the Tera-Promise and Jira repositories. We performed rigorous preprocessing and measured the results using metrics like MSE, RMSE, and AAE. The Hybrid Model is consistently superior, and especially with the MSE and AAE tasks that it is put through. This is a direct result of its high-level adaptability and strength. Such research radically changes the way software quality is improved by plugging the gaps associated with defect density prediction mechanisms and pushes the Hybrid Model as a viable machine learning approach.