Unformatted text preview: Moving Average model weighted average of past news (noise/errors) the covariance between the same lags is the same throughout in a time series f(x1, x2, x3) = f(xt+1, xt+2, xt+3) the covariance between x1 and x3 is the same as xt+1 and xt+3 i.i.d, the sample you get follows the same distribution and they are independent time series analysis → analyzing the data so we can take out the dependent residuals when you have a random walk, the variance increases over time dependent because tomorrow depends on where you are today compute a difference (gets rid of the dependency) D(Xt) = Xt - X(t-1) = rt …. rt is independent the variability of D(xt) is constant...
View Full Document
- Fall '13
- Variance, WI, Autoregressive moving average model, Autoregressive model