Trevor Trinkino, a quantitative analysts and trader at Kershner Trading Group recently put together an introduction to Machine Learning utilizing CloudQuant and Jupyter Notebooks. In this video he walks you through a high-level process for implementing machine learning into a trading algorithm, using CloudQuant’s trading strategy online tools and the Numpy, Pandas and SKLearn python libraries. The focus is on ensemble methods and utilizing SKLearn’s Random Forest Classifier to create a stronger alpha signal during a reversal setup.
He begins with pulling and cleaning data then goes through the process of implementing and tuning a Random Forest Classifier to improve the financial returns by a substantial margin. It shows you how to tune your classifier’s hyper parameters to help it learn more effectively and also covers how to choose your most important features to help the model perform better out of sample.
Mr. Trinkino graduated from the University of Colorado at Boulder with a Bachelors in Mathematics and Quantitative Economics, as well as a certificate in Actuarial Studies and Quantitative Finance.