To introduce a variety of statistical models for time series and cover the main methods for analysing these models.
Review of Linear Models: Model building in time series analysis. AR, MA, ARMA and ARIMA model and model building by Box-Jenkins approach. Stationarity and invertibility conditions, ARIMA(p,d,q) model, estimation of parameters for AR, MA, ARMA and ARIMA processes, identification of processes with ACF PACF, Model order estimation techniques-AIC, AICC, BIC, EDC, FPE and forecasting.
Forms of non-stationarity in time series, Unit root: Dickey-Fuller, augmented Dickey-Fuller and Phillips-Perron tests. Panel data models: Balance and unbalance panel data, estimation in random effect and fixed effect models. ARCH and GARCH processes and models with ARCH, GARCH errors.
Multivariate time series processes and their properties: Vector autoregressive (VAR), vector moving average (VMA) and vector autoregressive moving average (VARMA) processes.
Non Linear Time Series Models: Non-linear auto regression, Threshold principle and threshold models. Amplitude-dependent exponential autoregressive (EXPAR), fractional autoregressive (FAR), Product Autoregressive Model (PAR), random coefficient autoregressive (RCAR), discrete state space auto regressive bilinear (BL), non-linear moving average, autoregressive models with conditional heteroskedasticity (ARCH).
Non-linear Least Square Prediction: Non-linear least square prediction in non-linear autoregressive, nonlinear moving average, Bilinear and random coefficient autoregressive models