All about naive Bayes Classifier Algorithm | Mathematics | Machine Learning | Data Science

It is a classification algorithm, which is extremely fast than other algorithms. If you are stuck with the large dataset then go with it. It will do it very effectively.

Actually, it is a collection of the classifier algorithm. It is a family of algorithms in which every single algorithm assumes that the presence of one feature is independent of the presence of other features.

Photo by Crissy Jarvis on Unsplash

But also the assumption assumed by the Naive Algo is not correct in the real world, so we can also get undesired results. Because it assumes every features independent of each other but there may be a case that the features are directly dependent on each other. But still, it works well in practice and we use it.

Bayes Theorem:

Bayes theorem is used to find the probability of an event occurring given that another event already occurred.

We can apply Bayes theorem like this...

Here,

    c = the target class

    x = dependent feature vector of size n

c = {y}

x = {x1, x2, x3, ... xn}

Working of NB Algo:

1. Create a frequency table of every single attribute in the table.

2. Now create a likelihood table now for every instance there will be a target value.

3. With the help of the likelihood table, use the Bayes theorem directly to calculate the desired results.

Advantages:

1. Extremely Fast.

2. Suitable for large datasets.

3. Work strongly in multiclass classification.

4. For numerical input variable, it assumes the normal distribution which is a strong assumption.

Disadvantages:

1. The outcomes from this algorithm not taken seriously because this algorithm assumes that all the features are independent of each other.