Naive Bayes Classifier

<aside> 💡 sklearn.naive_bayes

</aside>

For a given class variable and dependent feature vector through, the naive conditional independence assumption is given by:

$$ P(x_i∣y, x_1, ..., x+{i−1}, x_{i+1}, ..., x_m) = P(x_i∣y) $$

Types

• Gaussian NB: continuous features
• Bernoulli NB: binary features
• Multinomial NB: multinomial distribution, classification of word counts for text classification
• Categorical NB: categorical distribution
• Complement NB: class imbalance, CNB regularly outperforms MNB (often by a considerable margin) on text classification tasks.