GenLin Model Nugget
A GenLin model nugget represents the equations estimated by a GenLin node. They contain all of the information captured by the model, as well as information about the model structure and performance.
When you run a stream containing a GenLin model nugget, the node adds new fields whose contents depend on the nature of the target field:
- Flag target. Adds fields containing the predicted category and associated probability and the probabilities for each category. The names of the first two new fields are derived from the name of the output field being predicted, prefixed with $G- for the predicted category and $GP- for the associated probability. For example, for an output field named default, the new fields would be named $G-default and $GP-default. The latter two additional fields are named based on the values of the output field, prefixed by $GP-. For example, if the legal values of default are Yes and No, the new fields would be named $GP-Yes and $GP-No.
- Continuous target. Adds fields containing the predicted mean and standard error.
- Continuous target, representing number of events in a series of trials. Adds fields containing the predicted mean and standard error.
- Ordinal target. Adds fields containing the predicted category and associated probability for each value of the ordered set. The names of the fields are derived from the value of the ordered set being predicted, prefixed with $G- for the predicted category and $GP- for the associated probability.
Generating a Filter node. The Generate menu allows you to create a new Filter node to pass input fields based on the results of the model.
Predictor Importance
Optionally, a chart that indicates the relative importance of each predictor in estimating the model may also be displayed on the Model tab. Typically you will want to focus your modeling efforts on the predictors that matter most and consider dropping or ignoring those that matter least. Note this chart is only available if Calculate predictor importance is selected on the Analyze tab before generating the model. See the topic Predictor Importance for more information.