Quiz M1.02 - Question 3

M1. The Predictive Modeling Pipeline M1. Quiz M1.02
#1

From the cross_validation notebook:

"The output of cross_validate is a Python dictionary, which by default contains three entries:

(i) the time to train the model on the training data for each fold,
(ii) the time to predict with the model on the testing data for each fold,
(iii) the default score on the testing data for each fold."

Because the time to train is stored, I thought it could be considered that a score has been “computed” on the training data_set. But a “training time” is not considered a “score”.
I guess I read the question too fast, but in my mind, “score” didn’t necessarily meant “prediction_score”.
I will try being more precise about the vocabulary ^

#2

We try to keep consistent with the terminology. A score is obtained with model.score(X_train, y_train) or model.score(X_test, y_test) generally. We also expect a score to follow the rule “greater is better” (to the contrary to an error where “lower is better”). Time does not qualify there :slight_smile:

1 Like
#3

La même :slight_smile:
the time to train the model on the training data for each fold
=> une clé/valeur avec un tableau de 5 “valeurs” sur les données d’entrainement ! ce qui n’est un score effectivement :slight_smile: