Finish Kaggle titanic with over 80%

A week has gone and finally, I have nailed this problem. 80% is not high, but it lands me on 4% which is OK for me.

The book Hands-On machine learning with scikit learn and Tensorflow really helps a lot in the process. It helps me to standardize what I have to do, so I don’t need to think over when doing another ML project.

I have split my notebook into 3, and each focus on different area

Spend 80% of your time in engineering your features

https://www.kaggle.com/ymlai87416/titanic-score-over-80-part-1-data-preparation

Then spend you 5% of time choosing model https://www.kaggle.com/ymlai87416/titanic-score-over-80-part-2-model-selection

Then spend your cheap machine time to find a model which is good and not overfit. https://www.kaggle.com/ymlai87416/titanic-score-over-80-part-3-fine-tune