结合灰度预测特征与CNNs的信息服务体育成绩预测

针对大数据体育成绩预测存在精度较低的缺陷,提出一种结合灰度预测特征与CNNs的体育成绩预测算法.通过等维动态GOM模型提取灰度特征,并构建CNNs模型完成对体育成绩时间序列的回归与预测.以百米赛跑体育成绩为研究目标,完成了体育达标人数预测和体育成绩预测两个对比实验.结果表明,等维动态GOM模型以及相应的CNNs模型分别在达标人数和成绩预测中获得了最优的预测结果.提出的算法显著优于传统算法,分别在平均精度和极端数据中获得了更好的预测结果.

Sport performance prediction with information service based on gray scale prediction features and CNNs

In order to solve the problem of low accuracy in sport performance prediction with big data, a sport performance prediction algorithm based on both gray scale prediction and CNNs was proposed. The gray scale features were extracted with an equal dimension and dynamic GOM model, and a CNNs model was established for the regression and prediction of time series of sport performance. The 100-meter race was taken as research object, and two comparative experiments for qualified person number and sport performance prediction were completed. The results show that the equal dimension and dynamic GOM model as well as the CNNs model obtain the optimal predicted results in the prediction for qualified person number and sport performance. The as-proposed algorithm is obviously superior to the traditional algorithms, and obtains better prediction results in aspects of average accuracy and extreme data.