偏最小二乘在短跑运动员选材中的应用

Application of partial least squares in selecting sprinters

运动员选材问题是一个极其重要而又复杂的问题,就方法而言,目前尚无重大突破.在此运用一种新型的多元数据分析方法———偏最小二乘回归(PLS)方法,选取13岁维吾尔族男生,就其体质测试中的形态、机能、素质等有关数据加以分析,建立了短跑成绩的预测模型.对各项指标的贡献率进行分析与计算,并做了精度分析,为科学选材提供了定量依据.

As a number of factors are involved in selecting talents, which are contradictory to each other, it has been an issue of great concern in sports field to find a method that can be applied to evaluate athletes. In this paper, we use a method of Partial Least Squares to fit the forecasting models for the achievements of sprint. And make an analysis on the data of physique, function and quality got from the test of a group of 13-year-old Urgy boys. Every index factor proportion is taken into account. We also make some precision analysis.