基于BP神经网络模型的大学生体质综合测评

Research on physical fitness assessment of undergraduates based on artificial neural network model

为了解大学生体质健康状况,优选大学生体质相关的敏感指标,进一步调整健康干预措施。本研究参照《国家学生体质健康标准》要求,于2015年对南昌大学前湖校区所有在校的本科生进行体质测评。利用误差反向传播算法,分别建立男、女大学生体质人工神经网络模型。结果表明:男、女生神经网络模型的准确度较高,拟合程度均较好。其中,男生模型的前三位重要性体质指标为50m跑、1 000m跑和肺活量;女生模型的前三位重要性体质指标为800m跑、50m跑和肺活量。构建大学生体质的ANN模型切实可行,对于优化大学生体质健康评价指标以及健康决策具有重要意义。

Aimed at understanding the physical fitness level and characteristics of college students, selecting better physical fitness indices and further adjusting health intervention, this study Evaluated the physical fitness of college students on Qianhu campus of Nanchang University in 2015. In accordance with the《Na- tional student physical health standard》revised in 2014, the indices of the students physical fitness were tested and graded,and the physical fitness level of college students was assessed. An objective and stable Error Back Propagation （BP） artificial neural network （ANN） has been built. The ANN models for both male students and female students fit well and their accuracy is relatively good. The ANN model for male students indicates the three of most influential physical fitness indices are 50m sprint, 1000m running and vital capacity. The ANN model for female students indicates the most three influential physical fitness indi- ces are 800m running,50m sprint and vital capacity. It is practical to build ANN models of physical fitness in undergraduates ,which will make great contributions to optimizing the indices of physical fitness assess- ment and health decisions.