基于数据挖掘的体育运动即时数据自动化采集系统设计

Design of Automatic Sports Sports Data Acquisition System Based on Data Mining

针对体育运动数据采集的疲劳程度判断不准确问题,设计了一种数据挖掘的体育运动即时数据采集系统。首先采集人体运动时的心电和肌电生理信号,通过时域分析法分析心电和肌电生理信号的各特征指标与疲劳程度的关系,选择相关性大的指标进行特征提取;然后将提取特征输入至构建的基于LSTM神经网络的运动疲劳估计模型中进行疲劳程度分类检测。实验结果表明,传统的SVM算法的心电和肌电信号的疲劳检测准确率分别为86%和88%,提出的LSTM疲劳分类模型针对心电信号和肌电信号的疲劳检测准确率分别为96%和90%,可以看出,LSTM模型比SVM算法的疲劳检测准确率高出6%~10%,且对心电信号的检测结果浮动更小,具有更强的稳定性。由此说明,基于LSTM的疲劳分类模型能够明显提升基于运动数据对疲劳程度的检测准确性,在体育运动即时数据采集系统中进行广泛应用和推广,具备一定的可行性。

For the inaccuracy of fatigue judgment in sports data collection,a data mining real-time data collection system is designed.Firstly,the ECG and EMG signals during human movement are collected,and the relationship between the ECG and EMG signals and the fatigue degree is analyzed by time-domain analysis method for feature extraction.Then,the extrac-ted features are input into the constructed motor fatigue estimation model based on LSTM neural network for fatigue classifica-tion and detection.The experimental results show that the fatigue detection accuracy of ECG and EMG algorithm is 86%for the traditional SVM algorithm and 88%,respectively,and the proposed LSTM fatigue classification model is 96%and 90%,respectively.It can be seen that the LSTM model is 6%-10%higher than that of the SVM algorithm,and the detection re-sults of ECG signal are less floating and have stronger stability.Therefore,it shows that the fatigue classification model based on LSTM can significantly improve the detection accuracy of fatigue degree based on sports data,and can be widely used and popularized in the sports real-time data acquisition system,which has certain feasibility.