基于多元统计的DOTA2游戏中英雄的分析

The Analysis of Heroes in DOTA2 Games Based on Multivariate Statistics

摘要: Dota2是一个热门的moba类游戏，通过对游戏角色的控制，互相间配合取得游戏的胜利，深受游戏玩家喜欢。本文选取了游戏中的116个英雄，9项相关指标作为样本，利用SPSS统计软件对英雄数据进行统计分析。首先，使用系统聚类法分析各个指标间的关联程度，将指标分为两大类。然后，在不同类指标下分别对英雄进行k均值聚类，并分析了其中一些英雄的特点。最后，使用因子分析的方法，找到两个影响9个指标的主因子。

Abstract: Dota2 is a popular moba game, through the control of the game characters and cooperation with each other to win the game. It is popular with gamers. In this paper, we select 116 heroes in the game and 9 relevant indicators as samples and use SPSS to analyze the hero data. Firstly, hierar-chical cluster method is used to analyze the degree of correlation among indicators, then the indicators are divided into two categories. After that, heroes are clustered by k-means under different indexes meanwhile the characteristics of some of the heroes are analyzed. Finally, the method of factor analysis is used to find two main factors influencing 9 indicators.