Flag the Close Calls

A lab wants a person to double-check any test sample the model is unsure about.

Task: write close_calls(name, k, threshold). name is "iris", "wine" or "breast_cancer".

• Load the dataset as X and y, and split with train_test_split(X, y, test_size=0.25, random_state=0).
• Train a KNeighborsClassifier(n_neighbors=k) on the training rows.
• For every test row, take the highest of its predicted probabilities.
• Return the positions (0, 1, 2, ... counting rows of X_test in order) of the test rows whose highest probability is less than threshold, in increasing order.