How to get confidence of model prediction in python & sklearn.svm.SVC

Question

I have created a support vector model which makes a prediction via model.predict() . Is there a way to see the probability or confidence of this prediction?

def svc_training(X, y):
    # Create a support vector classifier
    clf = SVC(C=1)
    # Fit the classifier using the training data
    clf.fit(X, y)
    return clf

Answer 1

You need to enable probability parameter:

probability: boolean, optional (default=False):

Whether to enable probability estimates. This must be enabled prior to calling fit, and will slow down that method.

So, clf = SVC(C=1, probability=True) and you can reach confidence scores with

predictions = model.predict(predict_df)
clf.predict_proba(predict_df)

where predict_df is simply the dataframe you want to predict.

Answer 2

You can use SVC.predict_proba(X) But you need to set probability=True when you create a instance of SVC . You can read in doc of SVC.predict_proba(X) :

Compute probabilities of possible outcomes for samples in X. The model need to have probability information computed at training time: fit with attribute probability set to True .

from sklearn.svm import SVC

def svc_training(X, y):
    # Create a support vector classifier
    clf = SVC(C=1, probability=True)
    # Fit the classifier using the training data
    clf.fit(X, y)
    return clf

X = np.array([[-1, -1], [-2, -1], [1, 1], [2, 1]])
y = np.array([0, 0, 1, 1])
clf = svc_training(X,y)

print(clf.predict([[-0.8, -1]]))
# [0]

print(clf.predict_proba([[-0.8, -1]]))
# [[0.8656444 0.1343556]]
# probability of '0' is 0.8656444
# probability of '1' is 0.1343556

print(clf.predict([[1, , 1.5]]))
# [1]

print(clf.predict_proba([[1, 1.5]]))
# [[0.12552049 0.87447951]]
# probability of '0' is 0.12552049
# probability of '1' is 0.87447951