FINM 33150

Regression Analysis and Quant. Trading Strategies

Spring Quarter
Instructor: Brian Boonstra
Syllabus

Quantitative trading strategies, employing investment decisions based on model output, are a major component of business operations in the finance industry worldwide.  We will present the major components of these strategies as found in several asset classes (equities, futures, credit, FX, interest rates and energy). A large proportion of the models involved in quantitative strategies are expressible in terms of regressions.  We will cover most of the ways they are used, including practical tricks and considerations, and concentrating particularly on achieving trustworthy performance.  Mathematically, we will cover the computation of linear regressions with and without weights, in univariate and multivariate cases, having least squares or other objective functions.  Of the major computation technologies actively used by the finance industry (C/C++, Matlab, Java, R, VB/Excel, C\#, Python) we have chosen R and Python for numerical computation, with (very) light usage of Excel and with data coming from Quandl and some proprietary sources.