Generative or Naive Bayes model

Generative (Naive Bayes) model for classification is based on Bays' theorem. Here we fit a suitable probability distribution $P_i(x)$ to each class of data. Then we make decision for a new point for which the probability is maximum $\pi_i P_i(x)$, where $\pi_i$ is the relative frequency of i^th class in the training dataset.

For the probability distribution Gaussian distribution is usual choice, but we can choose other distributions suitable to our problem.

Notebooks

• Naive Bayes to classify UCI winery dataset

Resources

• https://scikit-learn.org/stable/modules/naive_bayes