摘要
采用L-M BP算法(LevenbergMarquardt BP算法),设计了神经网络训练测试工具,对阵形、防守位置敏感度、进攻位置敏感度和射门都采用了神经网络来进行数据拟合,利用神经网络存储离线采集的数据,实现未训练到的数据的正确输出。
Uses L-M the BP algorithm （LevenbergMarquardt the BP algorithm）, has designed the neural network training testing tool, to the lineup, defended the position sensitivity, the attack position sensitivity and shoots all to use the neural network to carry on the data fitting, the use neural network memory off-line gathering data, and using the neural network the data correct output which exuded ability realization not to train.
出处
《微计算机信息》
北大核心
2008年第28期305-306,241,共3页
Control & Automation
基金
基金项目:总装备部(编号不公开)
