Software Fault Prediction (SFP) is important in improving source code quality and reliability. This study investigates various combi- nations of software code metrics and CodeBERT methods, focusing on extracting meaningful insights from code and comments to cre- ate a comprehensive code embedding. Advanced code-embedding techniques with state-of-the-art CB model leverage both code and comments for SFP across 41 different datasets. The results show that proposed approach effectively identifies faulty source code, achieving an accuracy of 80.39% and Area Under Curve of 0.7458.

Dr. Lov Kumar is currently working as Assistant professor in the Department of Computer Engineering , NIT Kurukshetra . He received his Ph.D. in Computer Science and Engineering from NIT Rourkela, under the supervision of Prof. S. K. Rath. His current research interests are in the area of Mining Software Repositories, Software Analytics, and Social Media Analytics. His thesis is titled “Predicting Software Quality Parameters using Artificial Intelligence Techniques and Source Code Metrics”. He was a Faculty Member (at Thapar University) from Aug 2017 to Dec 2017 and BITS Pilani from Jan 2018 to Jan 2023. He has delivered over 60 invited talks, over 100 international refereed publications in international conferences and journals, and four published book chapter to his credit. He has won several other awards including the Young Scientist Award, Best Researcher Award, and best paper Award. He has a broad range of interests and hobbies. He loves to play cricket, read books, play chess, and solve Sudoku puzzles.