Predictive and Specialized Modeling • Neural Networks • Model Reports • Training and Validation Measures of Fit
Previous
 • 
Next
Training and Validation Measures of Fit
Measures of fit appear for the training and validation sets. See Figure 3.5.
Generalized RSquare
A measure that can be applied to general regression models. It is based on the likelihood function L and is scaled to have a maximum value of 1. The value is 1 for a perfect model, and 0 for a model no better than a constant model. The Generalized RSquare measure simplifies to the traditional RSquare for continuous normal responses in the standard least squares setting. Generalized RSquare is also known as the Nagelkerke or Craig and Uhler R2, which is a normalized version of Cox and Snell’s pseudo R2. See Nagelkerke (1991).
Entropy RSquare
Compares the log-likelihoods from the fitted model and the constant probability model. Appears only when the response is nominal or ordinal.
RSquare
Gives the RSquare for the model.
RMSE
Gives the root mean square error. When the response is nominal or ordinal, the differences are between 1 and p (the fitted probability for the response level that actually occurred).
Mean Abs Dev
The average of the absolute values of the differences between the response and the predicted response. When the response is nominal or ordinal, the differences are between 1 and p (the fitted probability for the response level that actually occurred).
Misclassification Rate
The rate for which the response category with the highest fitted probability is not the observed category. Appears only when the response is nominal or ordinal.
-LogLikelihood
Gives the negative of the log-likelihood. See the Fitting Linear Models book.
SSE
Gives the error sums of squares. Available only when the response is continuous.
Sum Freq
Gives the number of observations that are used. If you specified a Freq variable in the Neural launch window, Sum Freq gives the sum of the frequency column.
If there are multiple responses, fit statistics are given for each response, and an overall Generalized RSquare and negative Log-Likelihood is given.
Confusion Statistics
For nominal or ordinal responses, a Confusion Matrix report and Confusion Rates report is given. See Figure 3.5. The Confusion Matrix report shows a two-way classification of the actual response levels and the predicted response levels. For a categorical response, the predicted level is the one with the highest predicted probability. The Confusion Rates report is equal to the Confusion Matrix report, with the numbers divided by the row totals.

Previous
 • 
Next
Help created on 7/12/2018