How to use custom scoring function in sklearn cross_val_score

Question

I want to use Adjusted Rsquare in the cross_val_score function. I tried with make_scorer function but it is not working.

from sklearn.cross_validation import train_test_split
X_tr, X_test, y_tr, y_test = train_test_split(X, Y, test_size=0.2, random_state=0)

regression = LinearRegression(normalize=True)
from sklearn.metrics.scorer import make_scorer
from sklearn.metrics import r2_score
def adjusted_rsquare(y_true,y_pred):
    adjusted_r_squared = 1 - (1-r2_score(y_true, y_pred))*(len(y_pred)-1)/(len(y_pred)-X_test.shape[1]-1)
    return adjusted_r_squared

my_scorer = make_scorer(adjusted_rsquare, greater_is_better=True)
score = np.mean(cross_val_score(regression, X_tr, y_tr, scoring=my_scorer,cv=crossvalidation, n_jobs=1))

It is trowing an error:

IndexError: positional indexers are out-of-bounds

Is there any way to use my custom function i.e; adjusted_rsquare with cross_val_score?

Matthieu Brucher · Accepted Answer · 2018-12-19T12:40:24.253

6

adjusted_rsquare(X,Y) is a number, it's not a function, just create the scorer like this:

my_scorer = make_scorer(adjusted_rsquare, greater_is_better=True)

You also need to change the score function:

def adjusted_rsquare(y_true, y_pred, **kwargs):

That's the prototype that you should use. You compare the actual result to the result it should have been.

edited Dec 19 '18 at 12:40

answered Dec 19 '18 at 12:04

Matthieu Brucher

21,634
7
38
62

I have changed the code little bit and tried with what you have suggested but it is not working and throwing an error like positional indexes are out of bounds. – merkle Dec 19 '18 at 12:37
Added a comment on the wrong function declaration, you don't do any prediction inside the function, that's tackled by the model you want to score. – Matthieu Brucher Dec 19 '18 at 12:41
Could you please tell me how to get y_pred from cross_val_score? – merkle Dec 19 '18 at 12:44
That's the job of `cross_val_score` to give you y_true and y_pred. That's what you get in your function. You get them by definition. – Matthieu Brucher Dec 19 '18 at 12:45
I changed and tried again but still facing same issue. – merkle Dec 19 '18 at 12:55
1

I got it. Thanks for your help – merkle Dec 19 '18 at 13:05

How to use custom scoring function in sklearn cross_val_score

1 Answers1

Linked