Article Per Year (5 Year)
p-Index From 2020 - 2025
0
P-Index
This Author published in this journals
All Journal Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer
Muhammad Hafiz Azhar
Program Studi Teknik Informatika, Fakultas Ilmu Komputer, Universitas Brawijaya
Computer Science & IT Control & Systems Engineering Education Electrical & Electronics Engineering Engineering

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

Found 1 Documents
Search

 1 
Analisis Sentimen pada Ulasan Hotel dengan Fitur Score Representation dan Identifikasi Aspek pada Ulasan Menggunakan K-Modes Muhammad Hafiz Azhar; Putra Pandu Adikari; Yuita Arum Sari
Jurnal Pengembangan Teknologi Informasi dan Ilmu Komputer Vol 2 No 9 (2018): September 2018
Publisher : Fakultas Ilmu Komputer (FILKOM), Universitas Brawijaya

Show Abstract | Download Original | Original Source | Check in Google Scholar | Full PDF (946.476 KB)

Abstract

As the number of review is rising, there is a need to make a system that can do classify a review belong to which class, in this case there are positive and negative classes. Furthermore, we also need to know what aspect that commented in the review. In this research, sentiment analysis at aspect level, Bag of Nouns feature has been used for clustering to get aspect and sentiment classification with score representation feature to classify sentiment. With categorical attribute for Bag of Nouns feature, K-Modes is considered capable for clustering. In sentiment classification, score representation has been used for LVQ2 that can handle the correlation between attribute and also become alternative for another machine learning algorithm. Based on the evaluation with Silhouette Coefficient, the optimal number for clustering balanced data set is 7 and 5 for unbalanced data set. Based on the evaluation with precision, recall, and f1-score, the performance of the balanced data set are 89,2% for precision, 89,13% for recall, and 89,12% for f1-score. The evaluation for unbalanced data set are 87,38% for precision, 73,07% for recall, and 76,46% for f1-score. It can be concluded that score representation can be used for sentiment analysis.
Co-Authors Putra Pandu Adikari Yuita Arum Sari