考虑数据长短期特征的体育成绩预测方法  

Sports Performance Prediction Method Considering Long-Term and Short-Term Characteristics of Data

作　　者：徐婧 XU Jing(Department of Police Tactics,Fujian Police College,Fuzhou Fujian 350007,China)

机构地区：[1]福建警察学院警察战术系,福建福州350007

出　　处：《计算机仿真》2024年第12期331-336,共6页Computer Simulation

摘　　要：为解决当前体育成绩预测性能不高的问题,提出一种综合考虑数据长短期特征的多元特征感知网络,用于大学生体育成绩预测。首先基于记忆模块,构建感知短期特征信息的LSTM模块,并通过长期特征映射嵌入向量构建感知长期特征信息的embedding模块。然后基于多元特征感知网络挖掘大学生体育成绩的长期特征与短期特征信息,最后在200个真实样本数据进行验证。实验表明所提算法按照9:1、7:3和6:4三种不同比例区分训练集与测试集,在多项指标上具有优越性。To solve the problem of low performance in sports performance prediction,a multi-dimensional feature perception network that comprehensively considers the long and short-term characteristics of data is proposed for predicting sports performance of college students.Firstly,based on the memory module,the LSTM module perceiving short-term feature information was constructed,and the embedding module perceiving long-term feature information was constructed through the embedding vector of long-term feature mapping.Then,based on the multi-feature perception network,the long-term characteristics and short-term characteristics of college students'sports achievements were mined.And finally,200 real sample data were verified.The experiments show that the proposed algorithm can distinguish the training set from the test set in three different ratios:9:1,7:3 and 6:4,and has advantages in multiple indexes.

关 键 词：体育成绩预测 多元特征 神经网络 

分 类 号：G434[文化科学—教育学] TP183[文化科学—教育技术学]