摘要
近年来,神经网络发展迅猛,在人工智能、分类识别、预测分析,图像处理等方面都取得了不错的进展。目前应用最广泛的神经网络是BP神经网络,其是按照误差逆向传播来训练多层前馈网络,调整权值和阈值。本文以人的步幅特征(左右步长、步宽,步角)为输入,人体特征(身高,体重)为输出来训练神经网络,最终用matlab神经网络工具箱实现。结果表明,在正负10的误差范围内,该神经网络对人的身高、体重,年龄的预测准确率分别为96.67%,56.67%,70%。
In recent years,neural networks have developed rapidly,and have made good progress in artificial intelligence,classification,recognition,prediction and analysis,image processing,etc. At present,the most widely used neural network is BP neural network,which trains the multi-layer feedforward network according to the error reverse propagation and adjusts the weights and thresholds. In this paper,human step characteristics(left and right step length,step width,step angle) are used as input and human body characteristics(height,weight) as output to train the neural network. Finally,it is realized by Matlab neural network toolbox. The results show that within the error range of plus or minus 10,the predictive accuracy of the neural network for human height,weight and age is 96.67%,56.67% and 70%,respectively.
作者
张庆时
Zhang Qingshi(Criminal Investigation Police University of China,Shenyang 110035,China)
出处
《山东化工》
CAS
2019年第15期139-141,共3页
Shandong Chemical Industry
关键词
神经网络
预测
步幅特征
MATLAB
人体特征
neural network
prediction
step characteristics
matlab
human body characteristics