Time Series: Representation and Modelling

This lecture by the instructor covers the stochastic properties of time series, defining time series as a set of observations recorded at different times. It explains stationarity, autocovariance, and special classes of stochastic processes like MA and AR. The lecture delves into spectral density, digital filters, and estimation techniques such as Yule-Walker equations and forecasting for ARMA processes. It also discusses model identification, checking, and overfitting using information criteria. Long-memory covariance models, ARCH models, and state space models are explored, along with multivariate time series modelling and Vector Autoregressive framework.