决策树算法在乒乓球比赛中的应用  

Application of Decision Tree Algorithm in Table Tennis Match

作　　者：王杰[1] 虞丽娟[1] 张辉[2] 黄华勇[3] 

机构地区：[1]上海体育学院运动科学学院,上海200438 [2]上海体育学院体育教育训练学院,上海200438 [3]华为技术有限公司,广东深圳518129

出　　处：《计算机工程》2010年第24期272-274,共3页Computer Engineering

基　　金：国家体育总局科研基金资助项目(09B029)

摘　　要：结合属性均值和信息增益率对C4.5决策树算法进行改进,建立国内外优秀男子乒乓球运动员的比赛决策树模型,将其运用于比赛技战术诊断、评估和决策支持。实验结果表明,改进算法的建树精度较高,诊断结果较符合比赛实际;比赛评估能合理区分双方技战术实力。与专家经验比较,决策树生成的规则简便明确,能对比赛提供更准确的决策支持。This paper improves C4.5 decision tree algorithm based on mean value and information gain, builds the match decision tree model of foreign home and abroad men, and applies it to diagnosis, evaluation and decision support in table tennis match. Experimental results show that the accuracy of building tree algorithm is high, the diagnosis results conform to reality fairly, the algorithm can make a distinction between two players technique and tactic assessment reasonably. Compared with the expert experiences, the rules that the decision tree produced is simple and concise, and it can support more accuracy match decision support.

关 键 词：决策树算法 技战术诊断和评估 决策支持 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]