Select the analysis options for Trend Analysis

Stat > Time Series > Trend Analysis > Options

Add a custom title, specify prior parameter values, or specify weights.


In Title, enter your own custom title to replace the default title.

Perform a weighted average trend analysis

A weighted trend analysis includes fitting the same trend model to previous data in order to obtain an improved fit to the current data. The smoothed trend line combines the previous estimates and the new estimates in much the same way as exponential smoothing. This smoothing of the parameter estimates filters out some of the noise from the estimates in successive cycles.

If you supply estimates from a previous trend analysis fit, Minitab does a weighted trend analysis. If the weight for a specific estimate is α, Minitab estimates the new parameter by:

α p1 + (1 - α) p2 , where p1 is the parameter estimated from the current data and p2 is the prior estimate.

  • Prior parameter values: Enter the parameter estimates from the previous trend analysis in the order that they appear in the Session window or the graph.
  • Weights for blending priors with new estimates: You can enter weights between 0 and 1 for each new estimate. Enter the weights in the same order as the estimates. Default weights of 0.2 will be used for each parameter estimate if you don't enter any. If you do enter values, the number of values that you enter must be equal to the number of estimates.

Minitab generates a time series plot of the data, and a second time series plot that shows trend lines for three models. The Session window displays the parameters and accuracy measures for all three models.

By using this site you agree to the use of cookies for analytics and personalized content.  Read our policy