Prediction - Survival Model
Last updated
Last updated
Predict with Survival Model.
From a step that creates a Survival Model, you can access this feature from 'Add' (Plus) button. Select one of the 3 "Predict" menus.
With "Data" dropdown, select data to predict on from the following options.
Training - Get predicted values on training data.
Test - Get predicted values on test data.
Data Frame - Get predicted values on other data frame.
Select "Type of Prediction" from the following options.
Linear Predictor - lp - Risk Score on a log scale.
Risk Score - exp(lp) - Score for the risk that the event happens within a unit time for the subject.
Expected Number of Events - Number of events that would happen on average for a subject that has same condition as the subject during the time the subject has survived.
Status at Specified Time - Predicted status of the subject at the specified time.
Select "Type of Residual" from the following options.
Martingale
Deviance
Score
Schoenfeld
Schoenfeld (Scaled)
DFBETA
DFBETA (Scaled)
Click "Run" button to run the prediction.