Personalized ranking for video games based on online reviews: An S-Kano-TOPSIS method integrating requirement categories and public opinion

The rapid expansion of the video game market intensifies customers’ difficulty in selecting preference-aligned games. Although online reviews offer valuable insights, effectively leveraging this information remains challenging. To address this, we propose S-Kano-TOPSIS, a personalized ranking method for video games that integrates requirement categories and public opinion. First, BERTopic is used to extract customer requirements (CRs), and their performance is evaluated via sentiment analysis using a BW-CNN model. Then, SHAP is applied to quantify the influence of each CR on customer satisfaction. The Kano model is employed to adjust CR importance based on their influence patterns. Furthermore, to reflect real-world decision-making, we incorporate preference similarity by analyzing reviews of games similar to those the customer has played. Finally, TOPSIS is used to generate rankings tailored to individual needs. Experiments on 72,000 reviews from eight video games demonstrate that the proposed method surpasses baseline approaches across multiple evaluation metrics. These results suggest that S-Kano-TOPSIS offers a structured and quantifiable approach to personalized video game ranking.