中文说明：支持向量机SVM可以使用多分类的问题，本次的代码实现的就是这个功能，里面含有训练和测试数据，分别含有12类标签，也就是都分为12类，另外代码采用了神经网络中的one-versus-one的思想，将数据分类K（K-1）/2个子类，分别调用SVM方法实现
English Description：
Support vector machine SVM multi classification problem can be used, the code implementation is the function, which contains the training and testing data, containing, respectively 12 class label, also is divided into 12 categories and also code using the neural network one-versus-one thought, the number according to the classification of K (k-1) / 2 subclass, respectively, called SVM method