说到特征降维/选择的问题，大部分EDA的套路都是从model训练的loss来判断feature importance。其实有一个简单易行而且很有效的办法是在CV里面用做feature permutation，对原始特征shuffle得到shadow（也可以加一些噪音），在通过zscore比较两者差异来判断importance，不断遍历筛选。在ESLII中593页有提到这个办法。R里面有一个包Boruta可以做这件事，py也有：https://github.com/scikit-learn-contrib/boruta_py
BY Data Science Archive
The S&P 500 slumped 1.8% on Monday and Tuesday, thanks to China Evergrande, the Chinese property company that looks like it is ready to default on its more-than $300 billion in debt. Cries of the next Lehman Brothers—or maybe the next Silverado?—echoed through the canyons of Wall Street as investors prepared for the worst.
If riding a bucking bronco is your idea of fun, you’re going to love what the stock market has in store. Consider this past week’s ride a preview.The week’s action didn’t look like much, if you didn’t know better. The Dow Jones Industrial Average rose 213.12 points or 0.6%, while the S&P 500 advanced 0.5%, and the Nasdaq Composite ended little changed.
Data Science Archive from in