K-fold cross-validation in TensorFlow - Interactive Code Practice

Practice - 5 Tasks

1fill in blank

easy

Complete the code to import the KFold class from scikit-learn.

TensorFlow

from sklearn.model_selection import [1]

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

AGridSearchCV

Btrain_test_split

Ccross_val_score

DKFold

Attempts:

3 left

💡 Hint

Common Mistakes

Using train_test_split instead of KFold

Importing cross_val_score instead of KFold

The KFold class is used to split data into K folds for cross-validation.

2fill in blank

medium

Complete the code to create a KFold object with 5 splits and shuffling enabled.

TensorFlow

kf = KFold(n_splits=[1], shuffle=True, random_state=42)

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

A5

B10

C1

D3

Attempts:

3 left

💡 Hint

Common Mistakes

Using 1 split which disables cross-validation

Using too few splits like 3 which is less common

Setting n_splits=5 creates 5 folds for cross-validation.

3fill in blank

hard

Fix the error in the loop to correctly split data indices using KFold.

TensorFlow

for train_index, [1] in kf.split(X):

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

Aval_index

Btest_index

Csplit_index

Dindex

Attempts:

3 left

💡 Hint

Common Mistakes

Using 'val_index' which is not returned by kf.split

Using a single variable instead of two

The kf.split() method returns train and test indices for each fold.

4fill in blank

hard

Fill both blanks to create training and testing sets from data arrays X and y.

TensorFlow

X_train, y_train = X[[1]], y[[1]]
X_test, y_test = X[[2]], y[[2]]

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

Atrain_index

Btest_index

Cval_index

Dsplit_index

Attempts:

3 left

💡 Hint

Common Mistakes

Using the same index for both training and testing

Using undefined variables like val_index

Use train_index to select training data and test_index for testing data.

5fill in blank

hard

Fill all three blanks to compile accuracy scores for each fold using a TensorFlow model.

TensorFlow

accuracies = []
for train_idx, test_idx in kf.split(X):
    X_train, y_train = X[[1]], y[[1]]
    X_test, y_test = X[[2]], y[[2]]
    model = tf.keras.Sequential([
        tf.keras.layers.Dense(10, activation='relu', input_shape=(X.shape[1],)),
        tf.keras.layers.Dense(1, activation='sigmoid')
    ])
    model.compile(optimizer='adam', loss='binary_crossentropy', metrics=['accuracy'])
    model.fit(X_train, y_train, epochs=5, verbose=0)
    loss, acc = model.evaluate(X_test, y_test, verbose=0)
    accuracies.append(acc)
print(f"Average accuracy: {sum(accuracies)/len(accuracies):.2f}")

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

Atrain_idx

Btest_idx

Cval_idx

Dsplit_idx

Attempts:

3 left

💡 Hint

Common Mistakes

Mixing train and test indices

Appending loss instead of accuracy

Use train_idx for training data and test_idx for testing data. The accuracy is appended from the test evaluation.