Supervised Learning

What is Supervised Learning?
Let us go through this brief introduction in three parts:


To understand supervised learning let us first go through the terms:
‘Supervised’ means to '‘observe and direct the work of’’, whereas,
‘Learning’ means “the acquisition of knowledge or skills through study, experience, or being taught.

These terms when added together forms, ‘Supervised Learning’. Supervised Learning is the type of Machine Learning algorithms which map a function to predict an output based on a given set of an input-output pair.

To make it easy to understand, let’s take the following example:

Ram is a fresher searching for a job opportunity. And with luck, he had cracked an interview call from a reputed firm. On the day of the interview, Ram was seated in front of the interviewer. Interviewer placed 6 apples on the table. All similar in shape and size but varied in the shades of the color red.

The interviewer told ram the prices of 1st, 3rd and 5th apple, and asked him to approximate the price of 2nd, 4th, and 6th apple.


The prices of apples were as follows:
1st: 60/-
3rd: 40/-
5th: 20/-

Ram, in a moment or so, observed the relationship of the shades with the price. He noted, as the shades of the apples went from dark to light, the price goes down.
He went on to predict the price of the other apples by using an approximation, which is as follows;

Given, the apples are arranged in ascending order of shades(as already given), Ram wrote a function.
Price of apple = -(position of apple)x10 + 70.
He predicted, the prices as follows:
2nd: 50/-
4th: 30/-
6th: 10/-

The interviewer was impressed by his method, and he got the job.


Above example was demonstrates supervised learning in real life. Given a set of data, that maps on to each other and have some relation between them, either positive and negative can be used to predict the values of each other. This is known as supervised learning.

Supervised Learning is broadly classified into two types:

  • Classification: **: It is a Supervised Learning task where the output is having defined labels(discrete
    value). Example of which can be considered as the opposite of the above task, if we were given a
    price range marked in discrete bins and we had to predict which apple falls in which range.

  • Regression : **: It is a Supervised Learning task where the output is having continuous value. Like the above-given problem. If a machine learning algorithm is used to predict the values that can range infinitely, we use regression algorithms

There are many types of supervised learning algorithms, few of them are:

* Support Vector Machines
* linear regression
* logistic regression
* naive Bayes
* linear discriminant analysis
* decision trees
* k-nearest neighbor algorithm
* Neural Networks
* Similarity learning