Multiclass classification of white wine quality:
a) Using the data/winequality-white.csv file, perform some initial EDA on
the white wine data. Be sure to look at how many wines had a given quality score.
b) Build a pipeline to standardize the data and fit a multiclass logistic regression
model. Pass multi_class='multinomial' and max_iter=1000 to the
LogisticRegression constructor.
c) Look at the classification report for your model.
d) Create a confusion matrix using the confusion_matrix_visual() function
from the ml_utils.classification module. This will work as is for
multiclass classification problems.
e) Extend the plot_roc() function to work for multiple class labels. To do so,
you will need to create a ROC curve for each class label (which are quality scores
here), where a true positive is correctly predicting that quality score and a false
positive is predicting any other quality score. Note that ml_utils has a function
for this, but try to build your own implementation.
f) Extend the plot_pr_curve() function to work for multiple class labels
by following a similar method to part e). However, give each class its own
subplot. Note that ml_utils has a function for this, but try to build your own
implementation