Homework 8

Tasks

Part 1

Use the data in data_clean.csv from this link

1.1

  • Divide the sample into train and test subsumples of 192 and 8 observations respectively.

  • select VAR orders p_aic and p_bic for the training sample according to AIC and BIC

  • Estimate VAR§ for p=p_aic and p=p_bic

  • Forecast variable ‘a’ 1 to 8 steps ahead with each VAR model

1.2

  • select ARMA(p,q) orders for variable ‘a’ using AIC and BIC (again for the training sample)

  • Estimate the selected ARMA(p,q) models using training sample.

  • Forecast 1 to 8 steps ahead with each ARMA model

1.3

Make a plot with variable ‘a’ showing the actual data for the full sample and the forecasts produced with each model.

Part 2

Use the data in sign_svar_data.txt from this link. Follow the instructions and complete the tasks in HW8-part2.ipynb

Submission

Submit as jupyter notebooks in the folder HW8 in your repository at https://github.com/PhDEcon108.