一、sklearn保存模型
https://www.cnblogs.com/zichun-zeng/p/4761602.html
二、python连接数据库
https://www.cnblogs.com/qixu/p/6133429.html
三、
arr1 = np.array(tp1,dtype=float)
四、python 逻辑回归
https://blog.csdn.net/java1573/article/details/78830463
五、NaN
检查数据中是否有缺失值:np.isnan(train).any()
np.where(np.isnan(df))
六、python写入csv
https://blog.csdn.net/waple_0820/article/details/70049953
https://www.cnblogs.com/yanglang/p/7126660.html
七、给矩阵添加一列
http://www.cnblogs.com/MrLJC/p/4823499.html