Course description
Most machine learning models focus on cross-sectional data, while most time-series models focus on time series with few variables and low-frequency data. This course covers the skills and models to handle big data that are both rich in variables and time. We discuss both structural models and reduced-form models. Students learn dynamic regression model, dynamic factor model, vector autoregressions model, error correction model, dimensional reduction tools for fat dataset, and state-space model. Students also learn advanced methods to decompose trend, cycle, and seasonality in high-frequency data and to make more reliable time series forecasting. For complete and current details about this Harvard Extension course, see the description in the DCE Course Search.