Connect with us

# Machine Learning: Supervised Learning Published

on  Machine Learning: Supervised Learning has a significant technique and it is the Naive Bayes algorithm. It is a method of creating classifiers, which are models that assign labels to instances based on certain features. Naive Bayes classifiers may also be called “Simple Bayes” or “Independence Bayes”. There are several different ways to build a Naive Bayes classifier, including probabilistic models, Gaussian, and multinomial models, and others.

A Naive Bayes Classifier is a program that deals with predictions of a class value given a set of attributes.

### For each known class value,

Once this has been done for all class values, output the class with the highest probability.

### P(h|d) = (P(d|h) * P(h)) / P(d)

Where:

You can see that we are interested in calculating the posterior probability of P(h|d) from the prior probability p(h) with P(D) and P(d|h).

For our example in this article, we’re going to focus on the probabilistic model. For this model, our instances are represented as a vector of independent variables called features. Based on this feature vector, probabilities for each class are assigned to each feature. For example, if we had the two classes “Male” and “Female”, the Boolean feature “beard” may have an 80% probability for men and a 1% probability for women. Thus, our one feature vector of (true) would be classified as “Male” with 80% confidence. These probabilities are trained by running instances through the classifier which we already have marked with the correct classes.

### The Naive Bayes “Machine Learning: Supervised Learning”

To better understand the Naive Bayes, Machine Learning: Supervised Learning we can say that, it is a classification algorithm for binary (two-class) and multi-class classification problems. This is a method, which deals with our understanding when we use binary or categorical input values. It is also called an idiot Bayes since the calculation of the probabilities for each hypothesis are shortened to make their calculation docile. Rather than attempting to calculate the values of each attribute value P(d1, d2, d3|h), they are supposed to be conditionally independent given the target value and calculated as P(d1|h) * P(d2|H) and so on.

This is a very strong assumption that is most unlikely in real data, i.e. that the attributes do not interact. Nevertheless, the approach performs surprisingly well on data where this assumption does not hold.

### Our Classifications Since our classifications are created using these feature vectors, it is important to make sure that your features are appropriate for your classes. For example, the feature “beard” is much more useful in classifying “Male” vs. “Female” than the feature “eye color” might be. It is also important to make sure your training data is an accurate representation of your instances. For example, if our training data happens to only have one male and one female, then we might end up in a situation where our classifier “thinks” 100% of females will have a beard, while no males will ever have a beard.  #### Latest posts by Atif Farid Mohammad (see all) Click to comment

1139173
Confirmed
61149
Deaths
236229
Recovered