Machine Learning for Finance
Winter Quarter
Instructor: Niels Nygaard
The course will focus on two Machine Learning categorization models: Logistic Regression and Support Vector Machines, both binary and multi-category. The course will develop the mathematical foundations for these models and the optimization algorithms for training them on actual data. The algorithms will be implemented in Python. The necessary parts of Python programming will be taught along the way as they are needed. The Machine Learning models will be used to train models for trading stocks based on both fundamental and technical data. The models will be implemented in Python, using several Machine Learning libraries such as Scikitlearn and back-tested using the web service Quantopian. At the end of the course, the students will develop and implement their own trading models and analyze the performance of their models.
*33160 is Python-coding intensive. A prerequisite for 33160 is Python: either via FINM 32500 or program approval. FINM students requesting approval in lieu of 32500 must submit requests by a deadline likely to be in September 2019, exact date to be announced.