Describe the workflow you want to enable

An enhancement to the output of confusion matrix function, better representing the true and predicted values for multilevel classes.

• i.e. Current Representation with code:
  from sklearn.metrics import confusion_matrix
y_true = ["cat", "ant", "cat", "cat", "ant", "bird"]
y_pred = ["ant", "ant", "cat", "cat", "ant", "cat"]
confusion_matrix(y_true, y_pred, labels=["ant", "bird", "cat"])
  Output:
  array([[2, 0, 0],
  [0, 0, 1],
  [1, 0, 2]])

Describe your proposed solution

When you have multiple levels you can have difficulty reading the ndarray, associating the levels with the True and Predicted values.

Proposed solution should look similar to the table below, providing better readability of the confusion matrix.

Possible Solutions:

1. Provide a parameter to prettyprint the matrix. printMatrix [type:bool]
2. Include another parameter to return ndarray, index as true_values, columns as predicted_values
   For example:
   cm = confusion_matrix(y_true, y_pred, labels=["ant", "bird", "cat"])
index=["true:ant", "true:bird", "true:cat"]
columns=["pred:ant", "pred:bird", "pred:cat"]
return cm, index, columns
   Which can be easily converted into a dataframe for further use