神经网络训练根据Kolmogorov定理，输入层有14个节点，所以中间层有29个节点

％中间层神经元的传递函数为 tansig

％输出层有8个节点，其神经元传递函数为logsig

％训练函数采用traingdx-neural network training under the Kolmogorov theorem, input layer has 14 nodes, Therefore, the intermediate layer has 29% of nodes middle layer neurons in the transfer function for the output layer tansig% have eight nodes, its neuron transfer function for the training function logsig% used traingdx