Ch 4 - ARMA processes Flashcards
What is an AR process?
The dependent variable is a function of p lags and a constant
What are the assumptions in an AR process?
The process is stationary, the innovation term is a white noise term
What do you mean by stationarity?
The unconditional mean and variance are same through time and the autocovariances of the data process is constant through time
What is a MA process?
Lags of the WN process error term
What is an ARMA process?
A combination of AR and MA models that enables us to model dynamic with fewer parameters
What is R^2?
R^2 measures the proportion of variation in the dependent variable explained by the model. We can also use R^2 to measure the degree of predictability in a time series process
Choosing between p and q involves a trade off between ______ and _______
estimation error and goodness of fit
What are some alternative measures of goodness of fit?
MSE, s^2, AIC, HQIC, BIC
What is in-sample overfitting?
Researchers add variables to a model that appear to be good because they increase the R^2 or lower the MSE, but they are not really good for forecasting