CODEWITHSUNDEEP
pythonclassificationprobabilitykeraspredict
Thiedent
Thiedent

Reputation: 126

Keras classifier predict_proba() doesn't match predict()

I'm working with Keras Theano backend on a classification problem with 14 classes. I want the predicted class plus the associated probabilities. The problem is the probabilities from predict_proba() don't appear to match the predicted class from predict(), here is the code plus the resulting output of 1 sample.

PPRANK = ['pp1', 'pp2', 'pp3', 'pp4', 'pp5', 'pp6', 'pp7', 'pp8', 'pp9', 'pp10', 'pp11', 'pp12', 'pp13', 'pp14', 'pp15']

FEATURES = (PPRANK)

# fix random seed for reproducibility
seed = 7
np.random.seed(seed)

data_df = pd.DataFrame.from_csv("data.csv")
X = np.array(data_df[FEATURES].values)
Y = (data_df["bres"].replace(14,13).values)


# define baseline model
def baseline_model():
    # create model
    model = Sequential()
    model.add(Dense(8, input_dim=(len(FEATURES)), init='normal', activation='relu'))
    model.add(Dense(14, init='normal', activation='softmax'))
    # Compile model
    model.compile(loss='sparse_categorical_crossentropy', optimizer='adam', metrics=['accuracy'])
    return model
#build model
estimator = KerasClassifier(build_fn=baseline_model, nb_epoch=200, batch_size=5, verbose=0)

#split train and test
X_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.1, random_state=seed)
estimator.fit(X_train, Y_train)

#get probabilities
predictions = estimator.predict_proba(X_test)

#convert expon to floats
probs = [[] for x in range(21)]
tick2 = 0
for i in range( len( predictions ) ):
    tick = 0
    for x in xrange(14):
        (predictions[i][(tick)]) = '%.4f' % (predictions[i][(tick)])
        probs[(tick2)].append((predictions[i][(tick)]))
        tick += 1
    tick2 += 1

# pprint probabilities
pp = pprint.PrettyPrinter(indent=0)
pp.pprint(probs)

#print class predictions
print estimator.predict(X_test)
print Y_test

Result:

#probabilities
[0.00000, 0.00030, 0.02360, 0.04329, 0.00019, 0.00069, 0.00120, 0.00030, 0.00559, 0.00410, 0.00510, 0.91549, 0.0, 0.0]
#predicted class
11
#actual class
13

It shows 12 having the highest probability from predict_proba() rather than 11 from predict().

Upvotes: 3

Views: 6386

Answers (1)

Alex I
Alex I

Reputation: 20287

The index of python arrays (and here classes) counts up from 0, not from 1. Have another look, 0.91 is the 12th value as people count things, but it's at index = 11 so predict and predict_proba are consistent.

As for why not 13, the prediction may very well be wrong (but check that you don't have the same kind of error there).

Upvotes: 4

Related Questions