关于机器学习简要知识点

1. 对于机器学习中特征工程的理解

参考文章：http://blog.itpub.net/29829936/viewspace-2648602/

1. 对于模型的理解 

引用自一篇博客https://www.cnblogs.com/bjwu/p/8970821.html

Features: numeric, categorical, ordinal, datetime, coordinate, text

Numeric features

All models are divided into tree-based model and non-tree-based model.

Tree-based models doesn’t depend on scaling      我的LightGBM属于树模型，不需要归一化 

Non-tree-based models hugely depend on scaling  WaveNet-神经网络需要

归一化问题属于特征工程范畴