Exercise

# K-nearest neighbors for mushrooms

The Gaussian Naive Bayes classifier did a really good job for being an initial model. Let's now build a new model to compare it against the Naive Bayes.

In this case, the algorithm to use is a 5-nearest neighbors classifier. As the dummy features create a high-dimensional dataset, use the **Ball Tree** algorithm to make the model faster. Let's see how this model performs!

Instructions

**100 XP**

- Build a
`KNeighborsClassifier`

with`5`

neighbors and`algorithm = 'ball_tree'`

(to expedite the processing). - Fit the model to the training data.
- Evaluate the performance on the test set using the accuracy score.