Naive Baye’s Algorithm
Hello readers , this blog explains about a machine learning algorithm called Naive Byes . We will deep dive into the basic concepts of Naive byes and understand how it works . This algorithm of machine learning considers some of the basic principles of probability for making predictions .
Basic Terminologies
Before going for the Naive Byes Algorithm , let us first understand some of the terminologies related to probability . For this , let us consider an example of a pack of cards which has 52 cards .
1) Independent Events
Let us consider we have two events —
- Drawing a Queen card from the pack of 52 cards .
- Draw a King card from the pack of 52 cards .
Now , let us try to find out the probability of these two events —
These two events do not rely on one other so they are not correlated and hence are called independent events .
2) Dependent Events
Now let us consider 2 more events as follows —
- Drawing a Queen card from the pack of 52 cards .
- Drawing another Queen card from the remaining pack of 51 cards .
Now , let us try to find the probability of these two events —
We can see that in above two cases , the second event is dependent on the first event . So , the second event is known as dependent event .
3) Conditional Probability
Now , if we consider the above Dependent Event 2 then we can present it in the form —
>>>>>>>>>>>>> P (event2 | event1) = 3/51 <<<<<<<<<<<<<<<<<
This says that the probability of Event 2 given that Event 1 has already occured is 3/51 . This is known as the conditional probability . Let us see the formula for conditional probability —
4) Baye’s Theorem
This is the Theorem that is used in the Naive Baye’s Algorithm for training and predictions .
Naive Baye’s Algorithm
Naive Baye’s is a supervised machine learning algorithm that works on the principle of Baye’s Theorem and is used for classification based learning and predictions . The most important assumption for Naive Baye’s is that any feature in the dataset is completely unrelated to any other feature in that dataset .
Let us try to understand the working of Naive Baye’s theorem from an example mathematically .
Let us consider a dataset in which we have 2 independent features Xa and Xb and one dependent feature Y . This Y is the output of out model and it has a binary output of Yes or No only . The dataset has ‘p’ records in all .
The probability of “Yes” or “No” in the Y is called “Prior” .
P (Yes) = (no. of Yes) / p
P (No) = (no. of No) / p
The Bayes Theorem for this dataset can be said as —
This is the Baye’s Theorem to be used in this dataset for training it over the data and doing predictions .
Now , since there are two outputs for Y i.e “Yes” and “No” , hence the Naive Baye’s algorithm finds the probability for both types of output as follows —
Now , to find the Output , the absolute Probability of “Yes” and “No” is calculated as follows —
Now , if the values of Pa(Yes) is Higher , then the output is Yes otherwise “No” .This is how the Naive Baye’s works in prediction .
That is all about this blog . Thank you for reading !!!