Which Column Is Lemon? — medium Machine Learning in Code problem | Incognition

Which Column Is Lemon?

10 pts · 10 coins

Machine Learning in CodeMedium

Which Column Is Lemon?

predict_proba gives one column per label, but which column is which? The model keeps the order in classes_, and it is not always the order you expect.

Task: write chance_of(train_rows, train_labels, new_rows, label, k).

Train a KNeighborsClassifier(n_neighbors=k) on train_rows and train_labels. The labels are strings.
For each row in new_rows, find the predicted probability of label. label is always one of the training labels.
Return those probabilities as a list, each rounded to 2 decimal places.

Examples

train_rows = [[150, 7], [170, 8], [130, 3], [110, 2], [160, 6]]

train_labels = ["lemon", "lemon", "apple", "apple", "lemon"]

new_rows = [[155, 7], [135, 4]]

label = "lemon"

k = 3

returns [1.0, 0.33]

train_rows = [[1, 1], [2, 2], [3, 3], [7, 7], [8, 8], [9, 9], [15, 15]]

train_labels = ["sun", "sun", "rain", "rain", "fog", "fog", "fog"]

new_rows = [[2, 2], [8, 8]]

label = "rain"

k = 4

returns [0.5, 0.5]

Hints

solution.py

1from sklearn.neighbors import KNeighborsClassifier
2
3
4def chance_of(train_rows, train_labels, new_rows, label, k):
5    pass
6