Analysis And Prediction Of Game Players Based On Data Mining Technology

With the development of technology,the game industry has gradually developed as an emerging industry.After the period of rapid development early,now,the current game player group has stabilized,and the competition in the game industry has become increasingly fierce.Successful games often require higher quality and users’ stickiness.Therefore,the analysis and of the player community is very important.This paper analyzes the active data of game players,including playing time and recharge amount,and try to establish a data mining model,and divide the game player into different groups.This paper also analyzes the characteristics of each player group.At the same time,this paper predict the player’s payment in future basing on the player’s current data.Based on the data of 1.01 million players in a game,we first analyzed the general characteristics of the player group from the perspective of revenue and payment.Then,using the K-Means clustering algorithm,with time-depth and paying-time-depth as the two division dimensions,we try to use clustering analysis on the player group.Finally the clustering results were analyzed in detail from the information about the total payment,activity,and the number of people who are in this group.Then we use deep learning which includes Gradient Boosting algorithm to build a model.This model can use the existing data of the player to predict whether the player will continue to pay for the game in the future,which means establishing a classification model of whether they will pay or not.Then we use regression analysis to predict Player’s payment amount,and obtained analysis results with commercial value.The research in this article provides ideas for game companies to analyze the player groups.Based on the ideas and results of the analysis,it is convenient for game companies to formulate future game marketing strategies and game design methods.