Just‐in‐time defect prediction enhanced by the joint method of line label fusion and file filtering

Just-In-Time (JIT) defect prediction aims to predict the proneness of software changes when they are initially submitted. It has become a hot topic in due its timely manner and traceability. Researchers have proposed many JIT approaches. However, these approaches cannot effectively utilise line labels representing added or removed lines ignore noise caused by defect-irrelevant files. Therefore, model enhanced joint method label Fusion file Filtering (JIT-FF) is proposed. Firstly, distinguish while preserving original information, authors represent code as original, added, codes according labels. Secondly, obtain semantics-enhanced representation, cross-attention-based fusion perform complementary feature enhancement Thirdly, generate containing fewer files, formalise filtering sequential decision problem propose reinforcement learning-based method. Finally, based on generated changes, CodeBERT-based commit representation multi-layer perceptron-based performed identify defective changes. The experiments demonstrate that JIT-FF can more effectively.