Use this to specify the results to be saved from a stepwise discriminant analysis in Genstat data structures.
- After selecting the appropriate boxes, you need to type the names for the identifiers of the data structures into the corresponding In: fields.
Save
| Selected model | Pointer | A pointer to the variates in the selected model (including initial terms) |
| Allocations to groups | Factor | Allocations of the units in the training set to groups |
| Steps in criterion from adding terms | Pointers | The criterion value for adding each term at each step of selecting the model. The values in each variate will be a missing value if that term has already been added to the model. |
| Unit scores for final model | Pointer or matrix | Of scores for units from the final selected model |
| Estimated error rate from model steps | Scalars | Estimated error rate (proportion of the units that were incorrectly allocated) for each step in selecting the model |
| Estimated specificity from model steps | Tables | Estimated specificity (proportion of the group that were correctly allocated) for each step in selecting the model |
| Loadings, roots and trace | lrv | Latent roots and vectors from the discriminant analysis for the selected model |
Display in spreadsheet
The saved results will be displayed within a new spreadsheet. Note, an LRV structure will have its component parts displayed in three separate spreadsheets.
See also
- Stepwise discriminant analysis menu
- Options for choosing which results to display
- SDISCRIMINATE procedure in command mode