Article Per Year (5 Year)
p-Index From 2020 - 2025
0.23
P-Index
This Author published in this journals
All Journal EDUMATIC: Jurnal Pendidikan Informatika
Maizaliyanti, Annisa
Unknown Affiliation
Computer Science & IT Education

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

Found 1 Documents
Search

 1 
Implementasi Algoritma Random Forest dalam Klasifikasi Ulasan Pengunjung Mall Semarang untuk Pengambilan Keputusan Layanan Maizaliyanti, Annisa; Umam, Khothibul; Yuniarti, Wenty Dwi; Handayani, Maya Rini
Jurnal Pendidikan Informatika (EDUMATIC) Vol 9 No 2 (2025): Edumatic: Jurnal Pendidikan Informatika
Publisher : Universitas Hamzanwadi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.29408/edumatic.v9i2.30379

Abstract

Visitor preferences for malls in Semarang are not optimal because bold reviews have not been utilized optimally in decision making. Our research aims to classify the sentiment of Google Maps reviews from 13 malls in Semarang with a total of 2,600 reviews. Labeling is done manually based on ratings, where ratings 1–3 are considered negative reviews and 4–5 as positive reviews. The classification method used is Random Forest because the ensemble approach (bagging) provides optimal results. The research process includes data collection, labeling, cleaning, data sharing, classification, and model evaluation. The data used is unbalanced and dominated by positive reviews, so the Synthetic Minority Over-sampling Technique (SMOTE) technique was applied. The overall accuracy before and after SMOTE remained the same at 84%. However, the model's performance in detecting negative reviews increased from 27% to 44% in recall and F1-score from 0.40 to 0.52, but these values ​​are still relatively low. Java Supermall Semarang is the mall with the best reviews, with a classification accuracy reaching 90%. This model is better at recognizing positive reviews, but less reliable for negative reviews. Therefore, its use as a decision-making preference needs to be done with caution. This research opens up opportunities for further development, including the use of other models such as BERT which are superior in understanding context and language in reviews.
Co-Authors Khothibul Umam Maya Rini Handayani Wenty Dwi Yuniarti, Wenty Dwi