摘要
针对人体足部尺寸预测中精度差、波动性大和研究方法不系统等问题,以122名女性,年龄(25.6±7.8)岁,身高(162.9±5.5)cm,体重(53.1±8.9)kg和84名男性,年龄(27.3±10.1)岁,身高(174±5.9)cm,体重(70.1±10.3)kg为研究对象,比较BP神经网络、GM(1,N)模型和多元线性回归等3种方法的预测性能。运用灰色关联分析理论,确定足长、足宽、前跗骨围及拇指高为跖围预测模型的基本部位,构建BP神经网络、GM(1,N)模型和多元线性回归等3种模型,并以平均绝对误差、均方根误差和平均相对误差等指标进行评价。结果显示:BP神经网络、GM(1,N)模型和多元线性回归的平均绝对误差分别为3.470%,4.725%和9.940%;均方根误差分别为4.494%,15.702%和10.813%;平均相对误差分别为1.474%,2.115%,3.967%。表明BP神经网络模型精度优于另外2种模型,更适合跖围尺寸的预测。
Due to the problems of poor accuracy,large fluctuation and unsystematic research methods in human foot size prediction,122 women(25.6±7.8 years old,height 162.9 cm±5.5 cm,weight 53.1 kg±8.9 kg)and 84 men(27.3±10.1 years old,height 174 cm±5.9 cm,weight 70.1 kg±10.3 kg)were selected as subjects,and the predictive performance is systematically compared by BP neural network,GM(1,N)models and multiple linear regression.The foot length,foot width,anterior tarsal circumference and thumb height,was determined as the measurements for plantar cicumference predication model by using grey correlation analysis theory,and then three relevant models were constructed and evaluated by the average absolute error,root mean square error and average relative error.The results show that BP neural network can be used to predict the foot length,foot width,anterior tarsal circumference and thumb height.The mean absolute errors of the three models are 3.470%,4.725%and 9.940%,respectively.The root mean square errors are 4.494%,15.702%and 10.813%respectively.The average relative errors are 1.474%,2.115%and 3.967%respectively.The accuracy of the BP neural network model is better than the other two models and is more suitable for the prediction of plantar circumference.
作者
李健
周捷
毛倩
马秋瑞
LI Jian;ZHOU Jie;MAO Qian;MA Qiurui(School of Apparel and Art Design, Xi′an Polytechnic University, Xi′an 710048, China)
出处
《西安工程大学学报》
CAS
2019年第6期595-601,共7页
Journal of Xi’an Polytechnic University
基金
陕西省科技厅计划项目(2018KW-056)
关键词
足型
跖围
BP神经网络
GM(1
N)模型
多元线性回归
foot type
plantar circumference
BP neural network
GM(1,N)mode
multiple linear regression
