A Novel Software Defect Prediction Method Based on Isolation Forest
Zhiguo Ding, Yuchang Mo, Zhusheng Pan · 2019
Software defect prediction (SDP) is a hot topic in the modern software engineering research community to analyze software quality and reliability. Many data mining and machine learning methods were used to judge whether software models are defective or non-defective by analyzing software source codes or developing processes, extracting some critical metrics and then building predictor. Isolation forest, being an anomaly detection method, was modified and introduced firstly in this application community, a novel SDP method based on isolation forest was proposed in this paper. Experiment conducted on five real NASA datasets demonstrated the proposed method effective.