Score


Scoring Models

To find detailed information on the specification for scoring models generated by a particular Algorithm, refer to the Algorithms section of the H2O documentation.

In general, score models by choosing the appropriate algorithm from the drop-down Score menu. Scoring requires the same format for testing data that was used for the training data.


Predicting

To leverage the model’s predictive capability, go to the drop-down Score menu and select Predict. A column for the predicted variable is not required, but the data should correspond to the training data in all other dimensions.

The prediction output is a single-column vector that corresponds row-for-row with the data set used for the prediction. H2O assigns a .hex key to the prediction that can be viewed by clicking the drop-down Admin menu and selecting Jobs.

To download the prediction output to the working directory as a .csv file, click the CSV link at the top of the Prediction Results page.