Scikit Learn - Classification with Naïve Bayes


Advertisements

Naïve Bayes methods are a set of supervised learning algorithms based on applying Bayes’ theorem with a strong assumption that all the predictors are independent to each other i.e. the presence of a feature in a class is independent to the presence of any other feature in the same class. This is naïve assumption that is why these methods are called Naïve Bayes methods.

Bayes theorem states the following relationship in order to find the posterior probability of class i.e. the probability of a label and some observed features, $P\left(\begin{array}{c} Y\arrowvert features\end{array}\right)$.

$$P\left(\begin{array}{c} Y\arrowvert features\end{array}\right)=\left(\frac{P\lgroup Y\rgroup P\left(\begin{array}{c} features\arrowvert Y\end{array}\right)}{P\left(\begin{array}{c} features\end{array}\right)}\right)$$

Here, $P\left(\begin{array}{c} Y\arrowvert features\end{array}\right)$ is the posterior probability of class.

$P\left(\begin{array}{c} Y\end{array}\right)$ is the prior probability of class.

$P\left(\begin{array}{c} features\arrowvert Y\end{array}\right)$ is the likelihood which is the probability of predictor given class.

$P\left(\begin{array}{c} features\end{array}\right)$ is the prior probability of predictor.

The Scikit-learn provides different naïve Bayes classifiers models namely Gaussian, Multinomial, Complement and Bernoulli. All of them differ mainly by the assumption they make regarding the distribution of 𝑷$P\left(\begin{array}{c} features\arrowvert Y\end{array}\right)$ i.e. the probability of predictor given class.

Sr.No Model & Description
1 Gaussian Naïve Bayes

Gaussian Naïve Bayes classifier assumes that the data from each label is drawn from a simple Gaussian distribution.

2 Multinomial Naïve Bayes

It assumes that the features are drawn from a simple Multinomial distribution.

3 Bernoulli Naïve Bayes

The assumption in this model is that the features binary (0s and 1s) in nature. An application of Bernoulli Naïve Bayes classification is Text classification with ‘bag of words’ model

4 Complement Naïve Bayes

It was designed to correct the severe assumptions made by Multinomial Bayes classifier. This kind of NB classifier is suitable for imbalanced data sets

Building Naïve Bayes Classifier

We can also apply Naïve Bayes classifier on Scikit-learn dataset. In the example below, we are applying GaussianNB and fitting the breast_cancer dataset of Scikit-leran.

Example

Import Sklearn
from sklearn.datasets import load_breast_cancer
from sklearn.model_selection import train_test_split
data = load_breast_cancer()
label_names = data['target_names']
labels = data['target']
feature_names = data['feature_names']
features = data['data']
   print(label_names)
   print(labels[0])
   print(feature_names[0])
   print(features[0])
train, test, train_labels, test_labels = train_test_split(
   features,labels,test_size = 0.40, random_state = 42
)
from sklearn.naive_bayes import GaussianNB
GNBclf = GaussianNB()
model = GNBclf.fit(train, train_labels)
preds = GNBclf.predict(test)
print(preds)

Output

[
   1 0 0 1 1 0 0 0 1 1 1 0 1 0 1 0 1 1 1 0 1 1 0 1 1 1 1
   1 1 0 1 1 1 1 1 1 0 1 0 1 1 0 1 1 1 1 1 1 1 1 0 0 1 1 
   1 1 1 0 0 1 1 0 0 1 1 1 0 0 1 1 0 0 1 0 1 1 1 1 1 1 0 
   1 1 0 0 0 0 0 1 1 1 1 1 1 1 1 0 0 1 0 0 1 0 0 1 1 1 0 
   1 1 0 1 1 0 0 0 1 1 1 0 0 1 1 0 1 0 0 1 1 0 0 0 1 1 1 
   0 1 1 0 0 1 0 1 1 0 1 0 0 1 1 1 1 1 1 1 0 0 1 1 1 1 1 
   1 1 1 1 1 1 1 0 1 1 1 0 1 1 0 1 1 1 1 1 1 0 0 0 1 1 0 
   1 0 1 1 1 1 0 1 1 0 1 1 1 0 1 0 0 1 1 1 1 1 1 1 1 0 1 
   1 1 1 1 0 1 0 0 1 1 0 1
]

The above output consists of a series of 0s and 1s which are basically the predicted values from tumor classes namely malignant and benign.

Advertisements