After following an introductory course of Machine Learning on Coursera and working with data scientists on commercial projects I wanted to apply Machine Learning to something practical myself.
I chose to apply it to Numerai. It's a hedge fund where trades are determined based on predictions crowdsourced from data scientists given anonymized data.
I created a 7 part series documenting my journey through the experiments:
- Part 1 - Basics
- Part 2 - First models: Random Forests, Support Vector Machine, Logistic Regression, Gradient Boosting, Deep Neural Network
- Part 3 - Automatic ML: auto-sklearn & Auto-WEKA
- Part 4 - Jim Fleming's models
- Part 5 - Automating with Luigi
- Part 6 - Evaluating predictions
- Part 7 - Concordant, original, 100% consistent