Article Per Year (5 Year)
p-Index From 2021 - 2026
0.408
P-Index
This Author published in this journals
All Journal Desimal: Jurnal Matematika Malcom: Indonesian Journal of Machine Learning and Computer Science
Mustafidah, Hilyatul
Unknown Affiliation
Computer Science & IT Education Mathematics Social Sciences

Published : 2 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 2 Documents
Search

 1 
Forecasting the Consumer Price Index in Bima City Using Random Forest Regression Naafia, Suci Faaza; Fathir; mustafidah, Hilyatul
Desimal: Jurnal Matematika Vol. 9 No. 2 (2026): Desimal
Publisher : Universitas Islam Negeri Raden Intan Lampung

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.24042/

Abstract

Accurate forecasting of the Consumer Price Index (CPI) is essential for supporting regional inflation monitoring and evidence-based economic decision-making. However, city-level CPI forecasting remains limited, particularly in smaller urban areas where local expenditure patterns may differ from national trends. This study aimed to develop a predictive model for the General CPI of Bima City, Indonesia, using Random Forest Regression based on monthly expenditure-group CPI data. A quantitative predictive research design was employed using secondary data published by the Central Statistics Agency (BPS) of Bima City from January 2013 to December 2025, comprising 156 monthly observations. The dataset was divided into training and testing subsets using an 80:20 ratio. Model performance was evaluated using Mean Absolute Error (MAE), Mean Squared Error (MSE), Root Mean Squared Error (RMSE), and the coefficient of determination (R²), while feature-importance analysis was performed to identify the most influential expenditure groups. The model achieved an MAE of 0.7626, an MSE of 1.4501, an RMSE of 1.2042, and an R² of 0.994, indicating strong predictive performance on the testing dataset. The actual–predicted comparison and scatter plot demonstrated close agreement between observed and predicted CPI values. Feature-importance analysis identified Finished Food, Health, and Foodstuffs as the dominant predictors contributing to model performance. These findings demonstrate that Random Forest Regression can effectively model the relationship between expenditure-group CPI variables and General CPI within the observed dataset while providing interpretable information regarding variable contribution. This study contributes to the growing application of interpretable machine learning in regional economic forecasting and offers a practical framework for supporting localized CPI monitoring and data-informed inflation analysis.
Komparasi Naive Bayes dan Support Vector Machine untuk Analisis Sentimen Kebijakan Pemerintah Mengenai Makan Bergizi Gratis: Comparing Naive Bayes and Support Vector Machine for Sentiment Analysis of Government Policy on Free Nutritious Meals Mulyani, Sri; Mutmainah, Siti; Mustafidah, Hilyatul
MALCOM: Indonesian Journal of Machine Learning and Computer Science Vol. 6 No. 4 (2026): MALCOM October 2026
Publisher : Institut Riset dan Publikasi Indonesia

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.57152/malcom.v6i4.2998

Abstract

Kebijakan Program Makan Bergizi Gratis (MBG) pemerintah memicu perdebatan publik di platform TikTok. Penelitian ini bertujuan mengklasifikasikan sentimen opini publik terhadap kebijakan MBG ke dalam kelas positif dan negatif dengan membandingkan algoritma Multinomial Naive Bayes dan Support Vector Machine (SVM) menggunakan ekstraksi fitur CountVectorizer. Dataset sebanyak 4.000 komentar dikumpulkan melalui Apify dari empat akun berita resmi nasional. Pelabelan awal dilakukan secara otomatis menggunakan InSet Lexicon dan divalidasi secara manual oleh anotator manusia sebagai gold standard untuk mengurangi bias makna kata informal. Hasil pelabelan menunjukkan dominasi sentimen positif sebesar 89,60% dan negatif sebesar 10,40%. Pengujian dengan rasio split data 80:20 menunjukkan bahwa SVM kernel linear menghasilkan akurasi tertinggi sebesar 99,12%, sedangkan Naive Bayes mencapai 91,62%. Pengujian lintas skenario menunjukkan Naive Bayes lebih rentan terhadap penurunan kinerja, dengan akurasi terendah 72,70% pada skenario undersampling dan tiga kelas, sementara SVM tetap stabil di atas 95%. Analisis WordCloud menunjukkan sentimen positif didominasi optimisme terhadap pemerintah, sedangkan sentimen negatif menyoroti risiko korupsi anggaran. Keterbatasan penelitian meliputi kesulitan model memahami konteks kalimat, kata informal, dan sarkasme. Hasil penelitian menunjukkan bahwa SVM lebih efektif dan stabil untuk klasifikasi sentimen teks TikTok yang tidak seimbang, memberikan kinerja konsisten pada berbagai skenario pengujian.
Co-Authors Fathir Fathir Naafia, Suci Faaza Siti Mutmainah Sri Mulyani