Tutorial 2

Question 1

What is meant by stationarity?

Answer 1

• for a weakly stationary time-series, its unconditional mean, unconditional variance and autocovariance at any lag do not change over time.
• the autocovariances are only functions of the lag and the parameters of the process

Question 2

Why is stationarity desirable for financial data?

Answer 2

• allows us to estimate the mean, variance and autocovariances from a single realisation of a time-series process and use these moments for forecasting

Question 3

What is ACF

Answer 3

• Autocorrelation Function
• measures the correlation between observations in a time series that are separated by a lag k

Question 4

What is PACF

Answer 4

• Partial Autocorrelation Function
• measures the correlation between y_t and y_t-k, controlling for the effects of intermediate lags.