The factors that influence how people play mobile games have been studied from a variety of perspectives in the wireless broadband environment. The original data in the background of the game, such as user operation records, consumption records, and social behavior records, are converted into user attributes, user tags are generated, and data sets are constructed in this study, which primarily uses data mining technology to study user behavior and form user portraits. By incorporating the similarity of players' subspace interests into the CFR (collaborative filtering recommendation) algorithm, a personalized game recommendation model, as well as the relationship management level of mobile game players, is created. The final fusion model's ROC-AUC value is 0.921, which has a percentile enhancement effect, according to the results. The findings show that using a personalized game recommendation model can help to improve the scalability of the CFR algorithm and the impact of data scarcity on the quality of mobile games recommended by players.
Read full abstract