Website-Based Liquid Selection Recommendation System Using Content-Based Filtering Method at Morevapor Gading Store

Abstract

Liquid is a favorite product among various vape lovers. This product provides a variety of unique and refreshing flavors, attracting the attention of vaping lovers to always try new variants. The high cost of purchasing vape liquid makes many people prefer to buy products recommended according to their preferences, making MoreVapor Gading the main choice. This study aims to develop a recommendation system for selecting vape liquid using a content-based filtering mechanism with the TF-IDF approach. The TF-IDF approach was chosen because of its ability to provide more precise weighting to relevant but not too common words, resulting in more accurate recommendations compared to other methods. In practice, the results of this study provide significant benefits for MoreVapor Gading, namely increasing the accuracy of product recommendations that can minimize ordering errors and increase customer satisfaction and loyalty. This research method uses a waterfall model consisting of the analysis, design, implementation, and testing stages. The results of the study show that from 21 datasets, the system can provide five recommendations with the highest similarity values, namely Cair Grape 0.1445, American Winter Grape Candy Magic 0.1243, Paradewa Grape Athena 0.1151, American Winter Magic Fanta Float 0.0923, and Foom Breeze Series Guava 0.0918 based on user preferences. The recommendation system developed aims to provide accurate recommendations and in accordance with user preferences in choosing vape liquid.

Keywords

Liquid Vape ; Content Based Filtering ; TF-IDF