JGuizard
4/8/2016 - 3:34 PM

Classification with Dato

Classification with Dato

#Create Logistic classifier
classification_model = graphlab.logistic_classifier.create(train_data,
                                                      target = 'sentiment',
                                                      features=['word_count'],
                                                      validation_set=None)
                                                      

#Get the classifier coefficents
classification_model.coefficients
                                                      
#Make a prediction with probabilty (0,1)                                                     
classification_model.predict(train_data, output_type='probability')

#Make a prediction with margin (-inf,+inf)
classification_model.predict(train_data, output_type='margin')