Article Per Year (5 Year)
p-Index From 2021 - 2026
1.607
P-Index
This Author published in this journals
All Journal Jurnal Teknologi Informasi dan Ilmu Komputer Sebatik Jurnal Abdimas PHB : Jurnal Pengabdian Masyarakat Progresif Humanis Brainstorming JURNAL IT: Media Infromasi STMIK Handayani Makassar Building of Informatics, Technology and Science INFOKUM Journal of Artificial Intelligence and Digital Business
Azahari, Azahari
Unknown Affiliation
Automotive Engineering Computer Science & IT Control & Systems Engineering Economics, Econometrics & Finance Electrical & Electronics Engineering Engineering Mechanical Engineering Public Health

Published : 11 Documents Claim Missing Document
Claim Missing Document
Check
Articles

Found 11 Documents
Search

 1   2 
Customer Sentiment Analysis of E-Commerce Products Using the Naïve Bayes Method and Word Embedding Harpad, Bartolomius; Azahari, Azahari; Salmon, Salmon
Building of Informatics, Technology and Science (BITS) Vol 7 No 3 (2025): December 2025
Publisher : Forum Kerjasama Pendidikan Tinggi

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.47065/bits.v7i3.8879

Abstract

This study discusses customer sentiment analysis toward e-commerce products using the Naïve Bayes method combined with Word Embedding techniques to enhance the semantic understanding of Indonesian-language customer reviews. The research background is based on the rapid growth of e-commerce, which has created a strong need to understand consumer opinions through online reviews. The main challenge in sentiment analysis lies in the complexity of natural language, such as the use of informal words, abbreviations, and diverse emotional expressions. This study utilizes 40,607 Tokopedia customer reviews across five product categories with three sentiment labels (positive, neutral, and negative). The research stages include data collection, text preprocessing (case folding, tokenization, stopword removal, stemming, and slang normalization), feature representation using Word2Vec and FastText, and classification using Multinomial Naïve Bayes. Experimental results show that the combination of Word2Vec and Naïve Bayes achieved an accuracy of 87.92%, while FastText and Naïve Bayes improved accuracy to 91.52%. The FastText-based model proved superior in handling morphological variations and non-standard spellings, making it more effective for Indonesian customer review texts. The WordCloud visualization reveals the dominance of positive words such as “sesuai” (appropriate), “barang” (item), and “cepat” (fast), indicating customer satisfaction regarding product conformity and service speed. The Confusion Matrix results indicate a bias toward the positive class due to data imbalance, where the model still struggles to recognize neutral and negative classes. Overall, this study demonstrates that integrating Word Embedding with Naïve Bayes enhances classification performance and provides richer semantic representations compared to traditional Bag of Words approaches. This approach has the potential to be applied in developing data-driven recommendation systems and marketing strategies within Indonesia’s e-commerce ecosystem.
Co-Authors Amelia Yusnita Ar Rahman, Nurbalingga Suhendra Ekawati, Hanifah Halim, Ryan Artanto Harianto, Kusno Harpad, Bartolomius Pahrudin, Pajar Pratiwi, Heny Rosita, Dewi Salmon Salmon Salmon Salmon, Salmon Sari, Aulia Rahmita Syamsuddin Mallala Ukkas, Irwan Yonatan, Kevin Yulindawati, Yulindawati Yunita Yunita