ECN7510 Financial and Macroeconomics Data Analysis
3 Elective Credits
There are numerous challenges to competently modeling/forecasting financial and macroeconomic variables. Basic statistical analysis may yield misleading results for various reasons including time-varying volatility, structural change, and outliers. Time series data is also often “non-stationary” which, if not addressed, can lead to spurious regression or “nonsense correlations”. Analysis of “big data” meanwhile can promote model selection bias where the effects in the final model are overstated. Moreover, the relationship of cause and effect between variables is often two-way, where both variables affect each other. In this class, you will learn several methods to address these problems including: cointegration (to address non-stationarity), vector autoregression (to allow for interdependence between variables), and a machine learning algorithm to conduct unbiased model selection from “big data” while addressing outliers and structural change.
Prerequisites: None
ECN7510: Financial and Macroeconomic Data Analysis
Academic Unit