神经网络SVM实现分类，采用高斯核，标准差经过试验，最终定在0.81。训练和测试样本在1到1000之间间隔取点，训练样本取奇数，测试样本取偶数，没有噪声-SVM neural network to realize classification, USES the gaussian kernel, the standard deviation after test, final set at 0.81.Training and testing samples in the interval bet