Following Jason Brownlee's machine learning tutorial since watching Good Tech Fest's Driven Data's tutorial on the flu shot uptake machine learning challenge. There are some categorical variables that are improving the model's prediction of flu shot uptake. That's what I want to dig into.
What's most important is NOT doing another 'Kaggle' submission. What's most important is to get a better understanding of WHICH variables are important in predicting flu-shot & H1N1 vaccination uptake.
Steps:
- complete JBL tutorial
- apply encoding to flu-shot categorical data
- will it allow me to eyeball the variables that DO improve the model?
- if not, what's next?
- consider whether to do a write up of the experience and share with Analytics Vidhya?