Classification Of Kredivo Application Reviews Based On User Satisfaction Aspects With The Svm Method

The development of the fintech sector in Indonesia has encouraged the creation of various digital payment applications, one of which is Kredivo which provides instant credit and installments without a credit card. In this study, we analyzed and classified Kredivo application user reviews based on satisfaction attributes using the Support Vector Machine (SVM) method. Review data was collected from the Google Play Store and pre-processed using text preprocessing, InSet dictionary-based sentiment tagging, TF-IDF feature extraction, and training-test data splitting in an 80:20 ratio. Based on the analysis, most Kredivo user reviews were observed to have positive sentiment of 38.70%, negative sentiment of 26.90%, and neutral of 34.40%. The SVM model developed for Kredivo review sentiment labeling works with positive, negative, and neutral. Word cloud visualization recognizes the most important words with positive tones such as "mantap", "baik", "cepat", "mudah", and "transaksi", as well as the most important words with negative tones such as "hapus", "bayar", "bulan", "meminjam", and "tidak". The results of this study can be feedback for Kredivo developers and other fintech platforms to improve services based on user needs and demands, as well as strengthen business strategies according to customer satisfaction levels.