Article Per Year (5 Year)
p-Index From 2021 - 2026
0.23
P-Index
This Author published in this journals
All Journal Jurnal IPTEK Bagi Masyarakat
Rahmatsyah Ananta Putra Ginting
Universitas Islam Negeri Sumatera Utara
Agriculture, Biological Sciences & Forestry Humanities Computer Science & IT Education Mechanical Engineering Physics Public Health Social Sciences

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

Found 1 Documents
Search

 1 
Decision Tree Classification of Rule-Based MSME Sector Labels for Widowed and Divorced Women Using Socioeconomic Factors Rahmatsyah Ananta Putra Ginting; Rakhmat Kurniawan
Jurnal IPTEK Bagi Masyarakat Vol 6 No 1 (2026)
Publisher : Ali Institute of Research and Publication

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.55537/j-ibm.v6i1.1820

Abstract

Widowed and divorced women may rely on micro, small, and medium enterprises to sustain household income, yet preliminary sector screening requires a consistent and transparent mechanism. This study developed a Decision Tree model to classify three rule-based MSME sector labels using demographic and economic characteristics. The dataset comprised 3,659 women aged 18–59 years in Kabanjahe District, Karo Regency, Indonesia. Age, occupation, marital-status category, monthly income, number of dependents, and business-ownership status were used as predictors. The target labels—culinary, fashion and handicrafts, and services and creative products—were constructed from predefined weighting rules. After preprocessing and a stratified 80:20 split, evaluation on 732 test records yielded 94.81% accuracy, 94.86% weighted precision, 94.81% weighted recall, and a 94.75% weighted F1-score. Low-income status had the highest impurity-based feature importance (43.11%), followed by the 18–25 age category (12.05%). The model consistently reproduced the constructed labels and produced interpretable decision rules. However, the results do not establish real-world sector suitability or business success because the labels were generated from the same predictor attributes rather than independent expert judgments or observed business outcomes.
Co-Authors Rakhmat Kurniawan R