Time Series Analytics and ForecastingJohn von Neumann Institute, Vietnam National University, Ho Chi Minh City Teaching Assistant: Hoai Nam Nguyen Monday 19/03/2018 - Thursday 29/03/2018 Total hours: 36 (including labs) (tentative) Timetable: First week: - Monday 19/03, h. 13.00-16.00 (class) - Tuesday 20/03, h. 8.30-11.30 (class+lab) - Wednesday 21/03, h. 13.00-16.00 (class+lab) - Thursday 22/03, h. 8.30-11.30 (class+lab) - Friday 23/03, h. 8.30-11.30 (class+lab), h. 13.00-16.00 (lab+lab) Second week: (hours have changed) - Monday 26/03, h. 13.00-16.00 (class+lab) - Tuesday 27/03, h. 13.00-16.00 (class+lab) - Wednesday 28/03, h. 8.30-11.30 (class), h. 13.00-16.00 (lab) - Thursday 29/03, h. 8.30-11.30 (tbd) - Friday 30/03, h 8.30-11.30 (student presentations) Useful links: - Central Statistics Office of Ireland (good source of time series data) - A little book of R for time series, by Avril Choglan - Meinhold and Singpurwalla (1983), "Understanding the Kalman filter" (pdf). MATERIAL (it will be uploaded day-by-day): Handouts (files are password protected): 19/03/18: handouts 1 (pdf), R example (R) 20/03/18: handouts 2 (pdf) 21/03/18: handouts 3 (pdf) 22/03/18: handouts 4 (pdf) 23/03/18: handouts 5 (pdf) 26/03/18: handouts 6 (pdf) 27/03/18: handouts 7 (pdf) 28/03/18: handouts 8 (pdf) Labs (files are password protected): 20/03/18: lab 1 (pdf,xlsx) and solutions (xlsx) 21/03/18: lab 2 (pdf,R), data (csv) and solutions (R) 22/03/18: lab 3 (pdf,xlsx) and solutions (xlsx) 23/03/18: lab 4 (pdf,xlsx) and solutions (xlsx); lab 5 (pdf,data) 26/03/18: lab 6: continue with lab 5 27/03/18: lab 7 (pdf,R) and solutions (R) 28/03/18: lab 8 (pdf,R,csv) and solutions (R) Other material 19/03/18: individual project instructions (pdf). Deadline for the submission is postponed to Monday 16/04/2018. 21/03/18: monthly beer production data (txt). Notice that in class we worked with a subsample of this dataset, starting Jan 1991. SYLLABUS (updated after each class): 19/03/18: introduction to quantitative forecasting; explanatory vs time series models; time series patters; simple linear regression to make forecasting. 20/03/18: autoregressive model AR(m); details of AR(1); transformation of the data to stabilise the variance; other transformations such as month length or trading day adjustment. 21/03/18: time series decomposition; additive (and multiplicative) model; regression trend and seasonal model (with AR component); regression trend and seasonal model (with indicator variables). 22/03/18: exponential smoothing; Holt's linear method; comparing different forecasting methods: RMSE, MAPE. 23/03/18: Holt Winters methods: additive and multiplicative; Stationarity in mean, autocorrelation function; partial autocorrelation function; differencing and seasonal differencing. 26/03/18: Backshift operator; moving average MA models; autoregressive moving average ARMA models; ACF and PACF in ARMA models; autoregressive integrated moving average ARIMA models; Aikake Inforation Criterion AIC. 27/03/18: ARIMA models with seasonality; Kalman-Filter KF model; derivation of KF updating equations. 28/03/18: GARCH models; GARCH and ARMA; one and k step ahead forecasting with GARCH models. |