Chapter 10 Flashcards
What is the definition of a Pseudo-true parameter? What does this parameter imply (name two)
What does LS minimize when it is incorrectly specified?
What is the definition of the Kullback-Leibler distance
What does ML minimize when it is incorrectly specified?
What are some tests to test if your model is correctly specified? On which idea is this based?
What is the Hausman specification test?
Which type of estimator do you use when you believe you have:
* a correctly specified model,
* outliers and fat tails,
* pre-sample information?
Why is the ML estimator the best estimator if your model is correctly specified?
How can Monte Carlo simulation be used to estimate parameter confidence intervals for small sample sizes?
How do we know which model is best for modeling a certain time-series?
What is the AIC?
What are some drawbacks of the AIC?
What is the main idea of the validation sample?
How can you find an approximate independent data set?
What is the main idea behind the Diebold-Mariano test statistic? What is the statistic?
The idea is to test the differences between FMAE or FMSE.
Why do you not always want to have a symmetric measure of analyzing forecasting accuracy?
Wh
Why is the Diebold-Mariano test not always useful? What are some drawbacks?
Name an alternative for sub-sample validation.
How can you backtest VaR?
What is the Kupiec’s VaR coverage test?
What are Ensamble Methods?
