pydata: Huiming's learning notes

Keep Looking, Don't Settle

动态规划

很早以前学运筹学的时候学过动态规划。最近看题目的时候看到这个名字,借助zhihu上的一篇文章,重新温习一下。

Data Engineering and Modeling 01: predict defaults with imbalanced data

This is an real question with a sample data from internet. We want to predict the defaults from the imbalanced data(default rate is about 0.08%). All the variables are hidden so we need to explore the variables attributes to find some realtion. Also it is imbalanced so we need to apply some sampling methods to balance the data.