ARMA 22\23 Flashcards
What is stationarity in time series?
A process is stationary if its mean, variance and autocovariance do no not change over time
How do you calculate the variance of a process et - iid(0,1)
Variance is the sum of the variances of all the terms
What is the defining feature of an AR(1) process in terms of autocorrelation
The autocorrelation decays exponentially
What is the purpose of an ACF?
It measures the correlation between a time series and its lagged values
What does the PACF indicate?
It shows the direct effect of a lagged variable after accounting for shorter lags
How does ACF differ for white noise vs AR(1)?
White noise has negligible auto correlation; AR(1) decays exponentially
What is the formula for an h-step ahead forecast in an AR(1) model?
Ø^h yt
What is the null hypothesis in the ADF test?
The process has a unit root (ø = 1)
What does rejecting the ADF test null hypothesis mean?
The process is stationary
What is the critical feature of an AR(1) process in PACF?
Sharp cutoff after lag 1
How is the mean of a stationary process related to time?
The mean is constant over time
What happens to autocovariance in a stationary process?
It depends only on the lag not on time t
Why is stationarity important in time series analysis>
Stationarity allows consistent estimation of model parameters
Why is PACF useful in empirical analysis?
It identifies the number of significant AR terms in a model
What is the definition of white noise in a time series?
White noise is a sequence of random variables with zero mean, constant variance, no autocorrelation