-, Muhammad Iqbal
IMPLEMENTASI ALGORITMA SUPPORT VECTOR MACHINE UNTUK ANALISA SENTIMEN DATA ULASAN APLIKASI PINJAMAN ONLINE DI GOOGLE PLAY STORE.
MALCOM: Indonesian Journal of Machine Learning and Computer Science.
ISSN 2797-2313
Abstract
Online lending (pinjol) have sparked significant controversy due to their easy access and widespread advertising on social media. Pinjol also employs distressing collection methods, high-interest rates, and short repayment periods, especially for illegal loans. Based on these issues, this study conducts a sentiment analysis on five pinjol applications: Kredivo, Easycash, Rupiah Cepat, Kredit Pintar, and Ada Pundi. Application review data was collected from Google Play Store using scraping techniques. Sentiment labeling was then performed automatically using the Indonesian sentiment dictionary (Inset). The labeling results indicate that all pinjol applications predominantly have negative sentiments. Kredivo has the highest number of positive sentiments (46%), while Easycash has the most negative sentiments (65%). The labeled data was then used for classification modeling with the SVM algorithm. Evaluation results show that the SVM algorithm performs quite well with an average accuracy of 72%, precision of 76%, and recall of 85%. However, specifically, SVM excels in classifying sentiments for the Kredit Pintar application with an accuracy of 83%. Visualization analysis using word cloud was also conducted to understand the context of user reviews of the pinjol applications. Observations show that users almost always discuss loan limits in every sentiment across all five applications.
Actions (login required)
 |
View Item |
