Prof. Govindan Rangarajan, IISc
Abstract: We start with univariate time series data and describe a popular representation of the data using autoregressive (AR) processes. Properties of AR processes will be considered in some detail. Next we generalize to multivariate time series analysis and their representation through vector autoregressive (VAR) processes. Finally we describe Granger causality that quantifies causal relations within a multivariate time series data.