摘要
为实现体育竞赛成绩的高效和快速预测,针对极限学习机预测结果易受其初始参数选择的影响的问题,结合果蝇优化算法的快速寻优能力,提出一种基于果蝇算法优化极限学习机的体育竞赛成绩预测模型(FOA-ELM)。选择第24届~第30届奥运会男子100m等7项田径赛冠军成绩为研究对象,进行滚动预测,研究结果表明,与PSO-ELM、GA-ELM、SOAELM、DE-ELM和ELM相比较,FOA-ELM具有更高的预测精度和计算效率,效果较好,具有较高的实际应用价值。
In order to achieve efficient and fast prediction of the athletic contest achievement,in view of the influence of extreme learning machine prediction results by its easy choice of initial parameters,combined with the rapid optimization ability of Drosophila optimization algorithm,this paper proposes an optimized extreme learning machine sports competition performance prediction algorithm(FOA-ELM)based on the model of Drosophila.of the twenty-fourth ~thirtieth Olympic Games,in cluding men's 100 mscores are seleded as the research object,using the rolling prediction method,is used The results show that,compared with PSO-ELM,GA-ELM,SOA-ELM,DE-ELM and ELM,the FOA-ELM has higher prediction accuracy and computational efficiency,the effect is good,and has better practical value.
作者
张文
牟艳
高振兴
刘志丰
Zhang Wen, Mu Yan, Gao Zhengxing, Liu Zhifeng(College of Internet of Things Engineering, Hohai University,Changzhou 21302)
出处
《微型电脑应用》
2018年第3期58-61,共4页
Microcomputer Applications
关键词
极限学习机
径向基神经网络
成绩预测
田径竞赛
extreme learning machine
neural network
radial basis function neural network
performance prediction track andfield-compction