Partial Least Squares

Predicted Response Tables - Prediction Interval

  

Minitab displays a prediction interval for each predicted response, providing a range within which you expect the predicted response for a single sample to fall.

The prediction interval is always wider than the confidence interval because of the added uncertainty involved in predicting an individual response.

Example Output

Predicted Response for New Observations Using Model for Moisture

Row      Fit    SE Fit        95% CI              95% PI

  1  14.5184  0.388841  (13.7343, 15.3026)  (12.5910, 16.4459)

  2   9.3049  0.372712  ( 8.5532, 10.0565)  ( 7.3904, 11.2193)

  3  14.1790  0.504606  (13.1614, 15.1966)  (12.1454, 16.2127)

  4  16.4477  0.559704  (15.3189, 17.5764)  (14.3562, 18.5391)

  5  15.1872  0.358044  (14.4652, 15.9093)  (13.2842, 17.0903)

  6   9.4639  0.485613  ( 8.4846, 10.4433)  ( 7.4492, 11.4787)

Test R-sq: 0.906451

 

Predicted Response for New Observations Using Model for Fat

Row      Fit    SE Fit        95% CI              95% PI

  1  18.7372  0.378459  (17.9740, 19.5004)  (16.8612, 20.6132)

  2  15.3782  0.362762  (14.6466, 16.1098)  (13.5149, 17.2415)

  3  20.7838  0.491134  (19.7933, 21.7743)  (18.8044, 22.7632)

  4  14.3684  0.544761  (13.2698, 15.4670)  (12.3328, 16.4040)

  5  16.6016  0.348485  (15.8988, 17.3044)  (14.7494, 18.4538)

  6  20.7471  0.472648  (19.7939, 21.7003)  (18.7861, 22.7080)

Test R-sq: 0.762701

Interpretation

In this example, the scientists requested a 95% prediction interval. They can be 95% confident that, for the first sample, the actual response for moisture will be between 12.5910 and 16.4459.