Try the
Default number of
lags first. Then, you can increase or decrease the number of lags after you examine the plot:
- Increase the number if the default misses a longer term pattern. However, usually the standard errors of the autocorrelations are reliable for only the first n/4 lags.
- Decrease the number to reduce the clutter in the plot.
In Number of
lags, enter the number (k) of past values that you want to correlate with each value in a time series. Minitab calculates the correlations for a lag of 1 through a lag of k.
For example, the table shows what a lag of 1 and a lag of 2 look like in the worksheet, but you do not need to create the lag columns. If you enter 2, Minitab calculates the correlations between the Variable column and Lag 1 and between the Variable column and Lag 2.
C1 |
C2 |
C3 |
Variable |
Lag 1 |
Lag 2 |
7.3 |
* |
* |
4.9 |
7.3 |
* |
6.5 |
4.9 |
7.3 |
3.8 |
6.5 |
4.9 |