ROC and AUC curves in TensorFlow - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to import the function that computes the ROC AUC score.

TensorFlow

from sklearn.metrics import [1]

Drag options to blanks, or click blank then click option'

Amean_squared_error

Broc_auc_score

Caccuracy_score

Dconfusion_matrix

Attempts:

3 left

💡 Hint

Common Mistakes

Importing accuracy_score instead of roc_auc_score

Importing confusion_matrix which is unrelated to ROC AUC

Importing mean_squared_error which is for regression

The roc_auc_score function calculates the area under the ROC curve, which measures how well a model distinguishes between classes.

2fill in blank

medium

Complete the code to compute the false positive rate and true positive rate for ROC curve.

TensorFlow

fpr, tpr, thresholds = roc_curve(y_true, [1])

Drag options to blanks, or click blank then click option'

Athresholds

By_true

Cy_pred

Dy_scores

Attempts:

3 left

💡 Hint

Common Mistakes

Passing true labels instead of predicted scores

Passing predicted class labels instead of scores

The roc_curve function requires the true labels and the predicted scores (probabilities) to compute the ROC points.

3fill in blank

hard

Fix the error in this TensorFlow code to compute AUC metric during model compilation.

TensorFlow

model.compile(optimizer='adam', loss='binary_crossentropy', metrics=[[1]])

Drag options to blanks, or click blank then click option'

Atf.keras.metrics.AUC()

Btf.keras.metrics.Accuracy()

Ctf.keras.metrics.Precision()

Dtf.keras.metrics.Recall()

Attempts:

3 left

💡 Hint

Common Mistakes

Using Accuracy metric instead of AUC

Using Precision or Recall which are different metrics

To track AUC during training in TensorFlow, use tf.keras.metrics.AUC() as a metric.

4fill in blank

hard

Fill both blanks to create a dictionary comprehension that maps thresholds to TPR values from ROC curve data.

TensorFlow

threshold_tpr = {thr[1] tpr[2] for thr, tpr in zip(thresholds, tpr)}

Drag options to blanks, or click blank then click option'

A:

B,

C.

D=

Attempts:

3 left

💡 Hint

Common Mistakes

Using comma instead of colon between key and value

Using equals sign instead of colon

Adding dot after variable names incorrectly

In dictionary comprehensions, key and value are separated by a colon :. To access the value of tpr, use the variable directly. The correct syntax is thr: tpr.

5fill in blank

hard

Fill all three blanks to create a TensorFlow callback that stops training when AUC reaches 0.95 or higher.

TensorFlow

early_stop = tf.keras.callbacks.EarlyStopping(monitor='[1]', patience=[2], mode='[3]')

Drag options to blanks, or click blank then click option'

Aauc

Bval_loss

Cval_auc

Daccuracy

E5

F10

Gmin

Hmax

Attempts:

3 left

💡 Hint

Common Mistakes

Monitoring 'val_loss' instead of 'val_auc'

Setting patience too high or too low

Using mode 'min' which stops when metric decreases

To stop training when validation AUC improves, monitor val_auc. Set patience to 5 to wait 5 epochs without improvement. Use mode max because higher AUC is better.