Explore Courses Blog Tutorials Interview Questions
0 votes
1 view
in Machine Learning by (47.6k points)

I would like to know if there is a way to implement the different score function from the scikit learn package like this one :

from sklearn.metrics import confusion_matrix

confusion_matrix(y_true, y_pred)

with tf.Session(config=tf.ConfigProto(log_device_placement=True)) as sess:

init = tf.initialize_all_variables()

for epoch in xrange(1):

        avg_cost = 0.

        total_batch = len(train_arrays) / batch_size

        for batch in range(total_batch):

       = {x:                          train_arrays, y: train_labels})

                avg_cost +=, feed_dict=                  {x: train_arrays, y:                                    train_labels})/total_batch

                if epoch % display_step == 0:

                     print "Epoch:", '%04d' %                               (epoch+1), "cost=", "                                   {:.9f}".format(avg_cost)

print "Optimization Finished!"

correct_prediction = tf.equal(tf.argmax(pred, 1), tf.argmax(y, 1))

# Calculate accuracy

accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))

print "Accuracy:", batch, accuracy.eval({x: test_arrays, y: test_labels})

Will I have to run the session again to get the prediction?

1 Answer

0 votes
by (33.1k points)

Formulas to calculate these metrics:


Implementation with Scikit learn:

pred = multilayer_perceptron(x, weights, biases)

    correct_prediction = tf.equal(tf.argmax(pred, 1), tf.argmax(y, 1))

    accuracy = tf.reduce_mean(tf.cast(correct_prediction, "float"))

    with tf.Session() as sess:

    init = tf.initialize_all_variables()

    for epoch in xrange(150):

            for i in xrange(total_batch):

           = {x: train_arrays, y: train_labels})

                    avg_cost +=, feed_dict={x: train_arrays, y: train_labels})/total_batch         

            if epoch % display_step == 0:

                    print "Epoch:", '%04d' % (epoch+1), "cost=", "{:.9f}".format(avg_cost)


    y_p = tf.argmax(pred, 1)

    val_accuracy, y_pred =[accuracy, y_p], feed_dict={x:test_arrays, y:test_label})

    print "validation accuracy:", val_accuracy

    y_true = np.argmax(test_label,1)

    print "Precision", sk.metrics.precision_score(y_true, y_pred)

    print "Recall", sk.metrics.recall_score(y_true, y_pred)

    print "f1_score", sk.metrics.f1_score(y_true, y_pred)

    print "confusion_matrix"

    print sk.metrics.confusion_matrix(y_true, y_pred)

    fpr, tpr, tresholds = sk.metrics.roc_curve(y_true, y_pred)

Hope this answer helps.

Welcome to Intellipaat Community. Get your technical queries answered by top developers!

28.4k questions

29.7k answers


94.1k users

Browse Categories