RECOMMENDATION THE BEST PRODUCT AT ONLINE MARKETPLACE (BUKALAPAK AND TOKOPEDIA) USING NAÏVE BAYES CLASSIFIER ALGORITHM

Abstract

Many products are sold on electronic commerce websites. As with Tokopedia or Bukalapak. Tokopedia is an online marketplace that every individual and business owner to open and manage their online store easily and cost-free, while providing a safe and convenient online buying and selling experience. Founded in 2009 by William Tanuwijaya, Tokopedia is rated as the market leader in the online marketplace in Indonesia. Bukalapak was established in 2010 as one of the digital agency products named Suitmedia, Bukalapak as one of the largest online products by Indonesian children in Indonesia. Achmad Zaky as CEO said there are currently about 150,000 sellers and 1.5 million active products in Bukalapak. Online customer reviews are available to assist in choosing the product to be purchased. Products must be recommended by special systems that are able to analyze and classify reviews because it is very difficult for users to read a lot of review, and the results of recommendations must be adjusted to the needs of users. This research is to classify customer reviews accurately into the opinion mining techniques of the recommendation system. This research focuses on classifying Indonesian language reviews to be positive or non-positive. In this study, we classify reviews on sentences and levels of review. This research succeeded in getting opinion mining techniques and better results in the review level classification than sentence-level classification. The results will contribute to the opinion mining phase of the recommendation system.