Venue: LSB LT1 (Wed 17:30-18:15)
Tutorial Notes & Schedule
Tutorial 0: Review of Preliminary Material
Notes (Without Solution) & (With Solution)
Tutorial 1: Decomposition of Time Series [Updated @20/09]
Notes (Without Solution) & (With Solution)
Handwritten notes (With Partial-Solution)
Tutorial 2: Stationarity and dependence measure
Notes (Without Solution) & (With Solution)
Handwritten notes (With Partial-Solution)
Tutorial 3: Parametric Time Series Models [Updated @28/10]
Notes (Without Solution) & (With Solution)
Handwritten notes (With Partial-Solution)
Tutorial 4: Midterm Review [Updated @29/10]
Notes (Without Solution) & (With Solution)
Handwritten notes (With Partial-Solution)
Tutorial 5: Inference on Model Parameters [Updated @11/11: New subsection added]
Notes (Without Solution) & (With Solution)
Handwritten notes (With Partial-Solution)
Tutorial 6: Model Selection and Diagnostics
Notes (Without Solution) & (With Solution)
Handwritten notes (With Partial-Solution)
Supplementary code (R-code)
Tutorial 7: Forecasting
Notes (Without Solution) & (With Solution)
Handwritten notes (With Partial-Solution)
Tutorial 8: Final Review
Summary Note
Tutorial 9: GARCH Model (Unexaminable)
Notes (With Solution)
Mock Final Exam (Fall 22-23)
(Question paper) & (Solution)
Useful link/reference
Argand plane (Visualization of the "unit circle" mentioned in tutorial 3): mathworld.wolfram.com/ArgandDiagram.html
General solution to linear recurrence relations (For solving explicit form of ACVF): https://brilliant.org/wiki/linear-recurrence-relations/
Partial fraction decomposition (For easing the algebra for evaluation of MA/AR representation): tutorial.math.lamar.edu/classes/alg/partialfractions.aspx
Errata list
Tutorial 1 (Updated @ 19/09)
In Exercise 6(c), the quantity of interest is Delta(Delta_9 X_20) instead of Delta(Delta_4 X_20)
Tutorial 3 (Updated @ 28/10)
In Figure 1, \Theta (Seasonal) was miswritten as \theta.
In P.2, the box "General guidance in Identification of Models", it should be p = deg(\phi) and q = deg(\theta) instead of p = deg(\theta) and q = deg(\phi).
For Exercise 6, gamma(0) = (9/25)sigma^2 instead of (25/9)sigma^2.
For Exercise 3(b), the conclusion should be Y_t = sum_{i=1}^\infty \psi_i Z_{t-i} instead of Z_t = sum_{i=1}^\infty \psi_i Y_{t-i}.
Tutorial 4 (Updated @ 29/10)
For Exercise 3, gamma(1) = 0 and rho(1) = 0. (It was not specified in the old version of solution)
For Exercise 4(c), the model should be specified as SARIMA(1,0,1) x (1,0,0)_2 instead as ARIMA(3,0,1).
Tutorial 5 (Updated @ 15/11)
Updated a new section (Section 5.5) regarding M-estimation.
For Exercise 8, assume {Z_t} ~{iid} N(0,\sigma^2) instead of WN(0,\sigma^2).
Tutorial 6 (Updated @ 15/11)
In P.1, the statement is sqrt{n} r_k -> N(0,1) instead of sqrt{n} C_k -> N(0,1).
Tutorial 7 (Updated @ 29/11)
In P.2, point 2 in the "Operation of Box and Jenkins Approach" box, Y_{n+k-i} is replaced by E(Y_{n+k-i} | Y_1,...,Y_n).
Feel free to email me through martinmtt@link.cuhk.edu.hk if you got any problem about the course or found any typo(s) in the tutorial notes.