Predicting Time Complexity Defects in Software Systems Using Machine Learning Techniques

This research presents a novel approach to predicting time complexity defects in software systems using machine learning techniques. We developed and evaluated multiple machine learning models to automatically identify potential performance bottlenecks and complexity issues in source code. Our approach achieved 87% accuracy in detecting time complexity defects across diverse code bases, demonstrating significant improvement over traditional static analysis methods. The resulting tool has been successfully integrated into common development environments, providing real-time feedback to developers and contributing to overall code quality improvement.