What question did this study set out to answer?

The goal is to enhance software defect prediction by addressing the challenges of imbalanced classes in data.

April 1, 2026Open Access

A Deep Learning Approach for Cost Effective Learning of Defective Prone Modules

Key Points

The goal is to enhance software defect prediction by addressing the challenges of imbalanced classes in data.
Proposed a deep learning approach for classifying imbalanced software data
Applied Principal Component Analysis for feature selection
Constructed a classifier using Adaptive Neuro Fuzzy Inference System
Evaluated classifier performance using Area under Curve (AuC) measures
Classifier performance improved compared to traditional neural networks
Demonstrated effectiveness in handling imbalanced and cost-effective data for software defect prediction

Abstract

Software Defect Prediction is a cost effective problem, in which the cost of majority class (Non defective) is low compared with the cost of minority class ( Defective). Learning from imbalanced data bias the classifier towards majority class. In this paper we are proposing a deep learning approach for classifying Imbalanced and Cost effective data. We applied Principle Component Analysis for feature selection and then constructed a classifier using Adaptive Neuro Fuzzy Inference System. The performance of the classifier was evaluated using AuC measures. We observed the performance of the classifier was improved compared with neural networks.

Bookmark

View Full Paper

Cite This Study

Maddipati et al. (Fri,) studied this question.

synapsesocial.com/papers/69cd79915652765b073a6724 https://doi.org/https://doi.org/10.14419/ijet.v7i4.16.28823

Bookmark

View Full Paper