Rekomendasi Toko Sepatu pada Marketplace Berdasarkan Aspect Based Sentiment Analysis pada Ulasan Produk Menggunakan Algoritma Support Vector Machine

Abstract

Along with the development of technology, buying shoes can be done through online shopping via the internet using a marketplace application. However, shopping through the internet also has several obstacles that can make buyers feel disappointed when shopping. One of the things that can help buyers in choosing stores and goods that are following their wishes is a review of the store's products. However, the large number of store product reviews will take time for buyers to read one by one. This study uses the Support Vector Machine method and the TF-IDF (term frequency-inverse document frequency) extraction feature to conduct aspect-based sentiment analysis on product reviews from the shoe store so that it can be used as a store recommendation for buyers who want to buy shoes. The data used comes from store product reviews taken from the Tokopedia application. Aspects to be studied include price, quality of goods, delivery of goods, and service. Sentiment analysis using the Support Vector Machine algorithm and the TF-IDF (term frequency-inverse document frequency) extraction feature provides sufficient accuracy of 86.87%.