Baseline
Evaluating a classifier relative to an appropriate baseline is important. This is especially true for imbalanced datasets, such as ad click-through, because high accuracy can easily be achieved through always selecting the majority class. In this exercise, you will simulate a baseline classifier that always predicts the majority class (non-click) and look at its confusion matrix, as well as what its precision and recall are.
X_train, y_train, X_test, y_test are available in your workspace. pandas as pd, numpy as np, and sklearn are also available in your workspace.
Este ejercicio forma parte del curso
Predicting CTR with Machine Learning in Python
Instrucciones del ejercicio
- Create
y_predan array of zeros with the same length asX_testusingnp.asarray(). - Print the resulting confusion matrix.
- Get the precision and recall scores.
Ejercicio interactivo práctico
Prueba este ejercicio y completa el código de muestra.
# Set up baseline predictions
y_pred = np.____([0 for x in range(len(X_test))])
# Look at confusion matrix
print("Confusion matrix: ")
print(____(y_test, y_pred))
# Check precision and recall
prec = ____(y_test, y_pred, average = 'weighted')
recall = ____(y_test, y_pred, average = 'weighted')
print("Precision: %s, Recall: %s" %(prec, recall))