长跑运动员耐力预测模型研究仿真  

Long Distance Athletes Endurance Simulation Prediction Model Research

作　　者：任海娟[1] 

机构地区：[1]郑州大学西亚斯国际学院,河南郑州451150

出　　处：《计算机仿真》2017年第5期305-308,共4页Computer Simulation

摘　　要：对长跑中运动员耐力进行准确预测,可以对运动员的身体变化进行合理分析,提出科学合理的训练指导意见。进行耐力预测时,需要考虑跑步中的加速度、天气等自变量,在这些自变量影响下提取运动员的运动负荷特征,以此为基础,建立耐力预测模型,但是传统方法很难有效表达这些因素产生的动态影响,无法提取准确的运动负荷特征,存在建模误差大的问题。提出基于改进人工鱼群神经网络算法的长跑中运动员耐力预测建模方法。先利用BP神经网络算法获取运动员在长跑中的踝关节加速度综合指数,利用计算的结果定义能耗预测中的自变量,提取长跑中运动员的运动负荷特征,利用人工鱼群算法对BP神经网络的初值权值和阈值进行优化,组建运动员耐力预测方程,利用支持向量机的默认RBF核函数搜索最佳的惩罚因子,利用搜索的结果对预测方差进行训练,获取最优的长跑中运动员耐力预测模型。仿真结果表明,所提算法可以为长跑中运动员耐力预测提供参考。A modeling method for endurance prediction of long distance athlete is proposed based on the modified artificial fish swarm neural network algorithm. Firstly, the BP neural network algorithm is used to obtain the acceler- ated speed comprehensive index of ankle joint. Then the calculation result is used to define the independent variables in energy consumption and the exercise load feature is extracted. The starter weight and threshold value of neural net- work are optimized by using artificial fish swarm algorithm to build athlete endurance prediction function. The default RBF kernel function of support vector machine is used to search the best penalty factor. Finally, the searching result is used to training the forecast variance and the optimization prediction model of athlete endurance in long distance running is obtained. The simulation results show that the proposed algorithm can provide the reference for the endur- ance prediction.

关 键 词：长跑 运动员 耐力预测模型 

分 类 号：G804.63[文化科学—运动人体科学]