因为在学ml过程中懒得重复完成实现算法以及分割数据集等这些琐碎工作，决定随便弄个模板以节省时间.
Steps: allocate -> train -> test
allocate:
给出train_set和test_set, 包括留出法，自助法，交叉验证法...
train(classification only for now):
binary real:
Linear Regression, Logistic Regression, LDA...
binary categorical:
Decision Tree...
multi-class wrapper:
OvO, OvR, MvM...
test:
测试指标设定，时间分析...