term 2 week 2 Flashcards
1
Q
what is serial correlation?
A
it is the presence of linear dependence over time
2
Q
what represents linear dependency over time?
A
the ACF - autocorrelation function
3
Q
what are the 4 models of ACF (autocorrelation function) ?
A
- white noise
- AR - autoregressive
- MA - Moving Average
- ARMA - autoregressive moving average
4
Q
what is white noise?
A
- weak stationarity: same mean, same variance, uncorrelated over time (no autocorrelation)
- weak dependence
- a random sequence of errors or shocks that exhibits no specific pattern or correlation over time
5
Q
what is homoskedasticity?
A
it means constant variance over time
6
Q
what does the white noise model tells us?
A
- if our residuals are white noise => the model has captured all systematic information; there is no pattern in the errors; the model is well specified
- under Gauss-Markov theorem OLS residuals are BLUE (best, linear, unbiased , estimators) => white noise, since uncorrelated and constant variance
7
Q
how does the ACF of whit noise look like?
A
it is perfectly flat since the association between today’s value and nay previous value is zero
8
Q
what is autoregressive model (AR) model?
A
9
Q
what Is an (MA) model?
A
10
Q
What is an ARMA model?
A
