Investigate a presented nonlinear time series (t=~5000) using at least three candidate models, including but not limited to SARIMA, analogue methods, persistance etc. Explain the rationale behind the model selection. Optimize these models (i.e. using AIC/BIC etc). Evaluate the forecast performance of the models using the Talagrand diagram and other statistical measures. Conclude the performance of the candidate models.
The report (including plots and other supporting diagrams) should be about 12-15 pages in length.
More details will be provided to the successful bidder.
This is an assignment for an online postgraduate programme which I am enrolled to. Hence, all work provided should be original (i.e. pass the plagiarism check). Payment will be withheld otherwise.
The deadline of this work is 1 May 2013.
I need the work to be of high quality. Hence, based on the overall grade of your project, I can award additional incentive money. This is to be negotiated.
The analysis should be completed with either R or Matlab. The codes used to generate the analysis should be sent to me. (i.e. It should be possible for me to replicate the analysis).
Only candidates with minimum qualifcations of MSc Statistics/Applied Mathematics/Econometrics or relevant should bid for the project.