I understand naive bayes completely until it’s variant comes multivariate , bernoulli and gaussian can anyone provide me a good blog or videos from where i can learn . I got the explanation on the web but all are in terms of text classification(bag of word model and tf-idf ) which i never used i am doing text classification using word embeddings so please suggest me some great blogs and videos from where i can learn thiese variants
You have three types of naive bayes - gaussian, binomial and multinomial. Lets see when can we use which one:
Gaussian - When the variables are continuous. You can use the gaussian naive bayes. A simple example would be that of the IRIS dataset.
Multinomial - When working with text based features, if you have the frequency of words in the corpus, you can use multinomial. For instance, if you have a sentiment analysis task.
Binomial - Instead of the frequency, if you have the binary features, denoting presence or absence of a word in the dataset, use binomial.