改进的SPRINT算法及其在体质数据分析中的应用  

Improved SPRINT Algorithm and Its Application in the Physical Data Analysis

作　　者：丁亚芝 郑志高[3] 马嵘[1] 

机构地区：[1]新疆师范大学运动人体科学重点实验室,新疆乌鲁木齐830054 [2]新疆师范大学体育学院,新疆乌鲁木齐830054 [3]北京大学软件与微电子学院,北京100260

出　　处：《体育科学》2014年第6期90-96,共7页China Sport Science

基　　金：教育部人文社会科学研究项目(11XJJC840001)

摘　　要：为了能够利用常规测试数据快速准确地确定人体体质状况,引入了纯区间的概念,并提出了一种趋势选择方法,在此基础上提出基于趋势选择的SPRINT算法(TESTSPRINT)。利用该算法可以根据身高、体重、握力等基本测试数据快速建立人体体质状况决策树,为快速确定人体体质健康状况提供依据。实验和理论分析结果均表明,提出的TESTSPRINT算法可以有效为分析人体体质健康状况提供决策依据,具有较高的准确性,并且时间和空间开销较原SPRINT算法小。In order to determine the human physical condition using the conventional tested data quickly and accurately, this paper proposed a trend selection based scalable parallelizable induc- tion of decision trees algorithm （TESTSPRINT）, based on the concept of pure interval and trend selection method. We can create a human physical condition decision tree quickly based on the basic test data such as height, weight and grip strength, according to the decision tree we can determine human physical health status quickly. Theoretic analysis and experimental dem- onstrations show that the algorithms proposed in this paper outperforms existing algorithms in time and space complexity, and it was proved fruitful applications in the decision human physi- cal health status with high accuracy.

关 键 词：SPRINT算法 基尼系数 体质数据 数据挖掘 

分 类 号：G804.49[文化科学—运动人体科学]