Random Forest — Business Insights

Draw Business Insights from RF 1. Var Imp: Look at the rank of important variables, if the top one are the least actionable variable, meaning that it’s impossible for company to change that variable, delete it and re-build RF check whether the top variable are continuous or categorical variable continuous variables tend to show up

Random Forest — Method and Application (Python)

Advantage of RF: Only little time is needed for optimization (the default param are good enough) Strong with outliers, correlated variables For continuous variables, it’s able to segmentize it Method: Create a bootstrapped dataset (Sample with replacement) Create a decision tree using the bootstrapped datasetBut only use a random subset of variables at each split

Linear Model in Python

Basic Work Flow: check distribution for each variable for assumptions check for outliers and any transformation need to perform (log?) fit a linear regression model Analyze model fitting (R square) predict (un-transform the result if transformation is performed)

EDA and Feature Engineering

Data Preparation Before landing a model for optimization or recommendation model, we need to make sure our data is in “ready-to-go” status. Here, I summarized some ways to clean data for future reference. Descriptive Stat Query and Merge Group and Plot Data fill na, replace and assign values Data Transformation on Column (Log, datetime) Check

Introduction

Hi there, okay this opening might be a little cheesy, but here we go. I always wanted to summarize my study, work and life in a way, so that when I look back I can see some mistakes I’ve already made, and achievements I’ve built. As you can see from the homepage title, it’s mainly