Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal Journal of Electronics, Electromedical Engineering, and Medical Informatics
Nur Hidayatullah, Wildan
Unknown Affiliation
Computer Science & IT Control & Systems Engineering Electrical & Electronics Engineering Engineering

Published : 1 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 1 Documents
Search

 1 
A Comparative Analysis of Polynomial-fit-SMOTE Variations with Tree-Based Classifiers on Software Defect Prediction Nur Hidayatullah, Wildan; Herteno, Rudy; Reza Faisal, Mohammad; Adi Nugroho, Radityo; Wahyu Saputro, Setyo; Akhtar, Zarif Bin
Journal of Electronics, Electromedical Engineering, and Medical Informatics Vol 6 No 3 (2024): July
Publisher : Department of Electromedical Engineering, POLTEKKES KEMENKES SURABAYA

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.35882/jeeemi.v6i3.455

Abstract

Software defects present a significant challenge to the reliability of software systems, often resulting in substantial economic losses. This study examines the efficacy of polynomial-fit SMOTE (pf-SMOTE) variants in combination with tree-based classifiers for software defect prediction, utilising the NASA Metrics Data Program (MDP) dataset. The research methodology involves partitioning the dataset into training and test subsets, applying pf-SMOTE oversampling, and evaluating classification performance using Decision Trees, Random Forests, and Extra Trees. Findings indicate that the combination of pf-SMOTE-star oversampling with Extra Tree classification achieves the highest average accuracy (90.91%) and AUC (95.67%) across 12 NASA MDP datasets. This demonstrates the potential of pf-SMOTE variants to enhance classification effectiveness. However, it is important to note that caution is warranted regarding potential biases introduced by synthetic data. These findings represent a significant advancement over previous research endeavors, underscoring the critical role of meticulous algorithm selection and dataset characteristics in optimizing classification outcomes. Noteworthy implications include advancements in software reliability and decision support for software project management. Future research may delve into synergies between pf-SMOTE variants and alternative classification methods, as well as explore the integration of hyperparameter tuning to further refine classification performance.
Co-Authors Akhtar, Zarif Bin Herteno, Rudy Radityo Adi Nugroho Reza Faisal, Mohammad Wahyu Saputro, Setyo