게임 구매자 리뷰 기반 신작 게임 추천 모델 연구

Abstract

As the untact lifestyle of modern people becomes normal due to the corona pandemic, the importance of leisure life continues to increase. The growth of the game industry, which has better accessibility and convenience, has increased remarkably. The future prospects of game industry are also expected to be large. Therefore, one of the main goals of game companies and game distributors is to settle users who have been brought on the platform. In this study, a method to build a personalized recommendation system based on reviews was presented. Users who wrote reviews on top-rated games were selected as recommended targets, and new popular games were provided as recommended items to induce the settlement of the platform. Game reviews, game information, and user information were collected from Steam platform, a global game distribution platform, and a list of top-rated games was collected from the Steam database. A review similarity-based recommendation model was created using Doc2Vec. By embedding the review using Doc2Vec, similar new popular game reviews were derived.