Investigation of Dataset Features for Just-in-Time Defect Prediction

TL;DR

This paper revisits the Kamei dataset for JIT defect prediction, highlighting preprocessing challenges, proposing new features for model training, and discussing dataset limitations affecting unsupervised learning.

Contribution

It identifies preprocessing issues, introduces new features for defect prediction, and analyzes the dataset's limitations for improved model development.

Findings

Preprocessing difficulties in the Kamei dataset

Proposed new features for defect prediction models

Limitations of the dataset for unsupervised learning

Abstract

Just-in-time (JIT) defect prediction refers to the technique of predicting whether a code change is defective. Many contributions have been made in this area through the excellent dataset by Kamei. In this paper, we revisit the dataset and highlight preprocessing difficulties with the dataset and the limitations of the dataset on unsupervised learning. Secondly, we propose certain features in the Kamei dataset that can be used for training models. Lastly, we discuss the limitations of the dataset's features.