Prediction
The functions listed below are all the methods made available by ERM for predicting y
and v
values on training and testing data. See the corresponding usage page to understand how to use these functions.
EmpiricalRiskMinimization.predict_v_from_test
— Function.predict_v_from_test(M [, theta])
Allows you compute unembedded predictions (i.e., in V space) based on a trained ERM model M on test data. Option to specify a choice of theta. It defaults to theta=thetaopt(M)
EmpiricalRiskMinimization.predict_v_from_train
— Function.predict_v_from_train(M [, theta])
Allows you compute unembedded predictions (i.e., in V space) based on a trained ERM model M on train data. Option to specify a choice of theta. It defaults to theta=thetaopt(M)
predict_y_from_u(M, U [, theta])
Allows you compute unembedded predictions (i.e., in V space) based on a trained ERM model M on one or many raw inputs, U
. Option to specify a choice of theta. It defaults to theta=thetaopt(M)
EmpiricalRiskMinimization.predict_y_from_test
— Function.predict_y_from_test(M [, theta])
Allows you compute embedded predictions (i.e., y values) based on a trained ERM model M on test data. Option to specify a choice of theta. It defaults to theta=thetaopt(M)
EmpiricalRiskMinimization.predict_y_from_train
— Function.predict_y_from_train(M [, theta])
Allows you compute embedded predictions (i.e., y values) based on a trained ERM model M on train data. Option to specify a choice of theta. It defaults to theta=thetaopt(M)
predict_y_from_u(M, U [, theta])
Allows you compute embedded predictions (i.e., y values) based on a trained ERM model M on one or many raw inputs, U
. Option to specify a choice of theta. It defaults to theta=thetaopt(M)
EmpiricalRiskMinimization.confusion_test
— Method.confusion_test(M)
Returns the confusion matrix based on test data for a trained ERM model
confusion_train(M)
Returns the confusion matrix based on training data for a trained ERM model