Initial Model Comparison Report shows an example of the initial Model Comparison report for a continuous response.
The Predictors report shows all responses and all models being compared for each response. The fitting platform that created the predictor column is also listed.
The Measures of Fit report shows measures of fit for each model. The columns are different for continuous and categorical responses.
The r-squared statistic. In data tables that contain no missing values, the r-squared statistics in the Model Comparison report and original models match. However, if there are any missing values, the r-squared statistics differ.
The root average squared error, the same value as RMSE except that RMSE adjusts for degrees of freedom.
One minus the ratio of the -log-likelihoods from the fitted model and the constant probability model. It ranges from 0 to 1.
A generalization of the Rsquare measure that simplifies to the regular Rsquare for continuous responses. Similar to the Entropy RSquare, but instead of using the log-likelihood, the Generalized RSquare uses the 2/n root of the likelihood. The maximum value is 1 for a perfect model. A value of 0 indicates that the model is no better than a constant model.
The average of -log(p), where p is the fitted probability associated with the event that occurred.
The root mean square error, adjusted for degrees of freedom. For categorical responses, the differences are between 1 and p (the fitted probability for the response level that actually occurred).
The average of the absolute values of the differences between the response and the predicted response. For categorical responses, the differences are between 1 and p (the fitted probability for the response level that actually occurred).
The rate for which the response category with the highest fitted probability is not the observed category.
Training and Validation Measures of Fit in Neural Networks provides more information about measures of fit for categorical responses.