Ø Two major objectives of this lab are to

o configure Python’s implementation of Naive Bayes classifier

o to evaluate these classifiers using a variety of different metrics.

Ø Configuring classifiers will be achieved using Python’s sklearn library. Evaluation will also be done via sklearn but use a dedicated set of methods designed specifically for computing metrics.

Same Dataset with Lab 7.

We had made some experiments about Decision Tree classifier in Lab 7.

• Configure the Naive Bayes classifier with the Gaussian option for numeric data

gnb = GaussianNB() #suitable for numeric features

gnb.fit(pred_train, np.ravel(tar_train,order='C')) predictions

= gnb.predict(pred_test)

print("Accuracy score of our model with Gaussian Naive

Bayes:", accuracy_score(tar_test, predictions))

• Configuring the Multinomial Naïve Bayes Classifier

By referring suitable information online identify the version of Naive Bayes suitable for classifying discrete data and fill in the line below.

mnb = MultinomialNB() #optimized for nominal features but can work for numeric ones as well

mnb.fit(pred_train, np.ravel(tar_train,order='C'))

predictions = mnb.predict(pred_test)

print("Accuracy score of our model with Multinomial Naive Bayes:", accuracy_score(tar_test, predictions))