Prediction of Software Defects using Ensemble Machine Learning Techniques

During software development and maintenance, predicting bugs becomes critical. Defect prediction early in the life cycle is an important aspect of quality assurance process that has received a lot attention previous two decades. Early detection defective modules can support team efficiently effectively utilizing available resources to provide high-quality products short amount time. The machine learning approach, which works by detecting hidden patterns among features, excellent way identify problematic modules. flaws NASA datasets MC1, MW1, KC3, PC4 are predicted using multiple classification algorithms this work. A new model was developed based on altering parameters XGBoost model, including N_estimator, rate, max depth, subsample. results were compared those obtained state-of-the-art models, our outperformed them across all datasets.