| Sample
|
You can use the options in this group box to generate
KNN results for the specific sample of your choice -
Examples,
Test, or
Overall - whenever applicable.
Note: For
K-Nearest Neighbors, it's not entirely meaningful to have predictions made on the example sample since the closest point in the feature space to an example is the example itself. This means
KNN predictions on an example set will always be identical to the outcomes of the set itself. Given this, selecting the
Examples option button will render some of the options on the
Quick tab unavailable, such as
Descriptive statistics,
Predictions,
Residuals, and
Confidence levels (or
Standard deviation for regression).
|
| Summary
|
Select this check box to produce a spreadsheet containing a brief summary of the current
KNN analysis.
|
| Model
|
Select this check box to create a spreadsheet containing the
KNN model specifications including the Prototype sample, number of nearest neighbors K, the distance measure, and the averaging scheme; refer to the Introductory Overview for details.
|
| Descriptive statistics
|
Select this check box to create a spreadsheet containing statistics relating to regression or classification results (whichever is appropriate). For regression tasks, the spreadsheet will contain the means and standard deviations of the observed and predicted variables, the S.D. ratio, and the correlation coefficient. For classification tasks, two spreadsheets containing the classification summary and a confusion matrix will be displayed.
|
| Cross validation error
|
Select this check box to produce the graph of K against the cross-validation error. This button is available only if you use cross-validation for estimating the number of nearest neighbors K.
|
| Predictions
|
Use the options in this group box to compute predicted values and residuals/accuracy for each dependent variable specified in the analysis. Depending on the choice of options in the
Include group box (see below), the results will include the values for the independent or predictor variables, dependent or outcome variables, predicted values, and/or residual (accuracy) values.
|
| Include
|
In this group box, select the type of values that you want to include in the results when displaying
Predictions and
Histograms (see below).
|
| Independents
|
Select this check box to include the independent variable values in spreadsheets and graphs.
|
| Dependents
|
Select this check box to include the dependent variable values in spreadsheets and graphs.
|
| Predictions
|
Select this check box to include predicted values in spreadsheets and graphs.
|
| Accuracy / Residuals
|
Select this check box to include prediction residuals for regression (Accuracy for classification) in spreadsheets and graphs.
|
| Confidence / Standard deviation
|
Select this check box to include classification
Confidence levels or standard deviation for predictions (regression tasks).
|
| Predictions
|
Select this check box to create a spreadsheet of the variables and results selected in the
Include group box.
|
| Histograms
|
Select this check box to create histograms of the variables and results selected in the Include group box.
|
| Save
|
Select this check box to display a standard variable selection dialog box, which is used to select variable(s) to be displayed together with other quantities of interest selected in the
Include (see above) group box.
Options / C / W. See Common Options.
|
| OK
|
Click the
OK button to accept all the specifications made in the dialog box and to close it. The analysis results will be placed in the
Reporting Documents node after running (updating) the project.
|
