Perform multiclass classification of white wine quality using k-NN and oversampling as follows:
a) Using the data/winequality-white.csv file, create a test and training set
with 85% of the data in the training set. Stratify on quality.
b) With imblearn, use the RandomOverSampler class to over-sample the
minority quality scores.
c) Build a pipeline to standardize data and run k-NN.
d) Run grid search on your pipeline with the over-sampled data on the search space
of your choosing to find the best value for k-NN's n_neighbors parameter
with scoring='f1_macro'.
e) Look at the classification report for your model.
f) Create a confusion matrix using the confusion_matrix_visual() function
from the ml_utils.classification module.
g) Plot a precision-recall curve for multiclass data using the plot_multiclass_
pr_curve() function from the ml_utils.classification module.