PROPOSAL FOR A GAME RECOMMENDATION SYSTEM BASED ON REVIEW KEYWORDS

초록

With the growth of e-commerce market, personalized recommendation sys­tems have emerged as a critical success factor in various industries. In the video game industry, previous recommendation systems were intended for existing users and required historical data on user purchases and ratings. In this study, a game recommendation system based on customer reviews was developed. It makes it possible to serve new users even when the data about them are limited. In the proposed system, games are represented in the form of keywords extracted from customer reviews. Recommendations are made based on keyword similarity. The KeyBERT and Word2vec techniques are applied for the natural-language processing of customer reviews. © 2023, ICIC International. All rights reserved.

키워드