What is machine learning: types, examples, definition and applications

Machine learning as the name suggests is the art of making the machine learn by itself by providing enough data to the machine. Google's prediction searches algorithm, Siri speech recognition system, and Netflix's recommendation system are all examples of machine learning. 

Machine learning can be broadly divided into three types 

    •  Supervised learning

    •  Unsupervised learning

    •  Reinforcement learning

So, What is Supervised Learning?  

In Supervised learning., we train the machine or model using labeled data. 

So what is labeled data? 

It's nothing but data that is already tagged with the right answers. We make the model to learn the patterns which it can use to predict the outcomes on your data. In some sense, supervised learning is nothing but the learning which happens under supervision. 

Let's take a real-world example. Suppose you have a fruit. Well, you'll be able to tell me it's an apple. But can the machines do that Well? Initially, no, because the missions do not know what will happen to look like. So let's give the machine thousands of images of an apple and make it learn from it. 

Well, it will be able to learn what an apple looks like, its size, its color, its texture, and other thousands of features. 

Now, the next time you give the image of an apple, the mission would be able to instantly recognize that it's an apple. The process by which you make the mission to learn from past label data and lose that learning to make an inference about the future data, that is called Supervised Learning.

Supervised learning can be further subdivided into two categories 

Regression - In Regression the type of variable which is going to predict will have a certain range of values. For example, the price of the house, the number of kilometers which have run today morning on a treatment. These can have any set of values. Such a type of problem is known as a regression problem.

Classification - Whereas in classification Variables you're trying to predict can help a certain set of discrete values. For example, whether to approve a customer's lawn, whether the customer will subscribe to a term deposit or not. These questions can have only two values are yes or no. Such a type of problem is known as classification problems.

The second type of Machine learning is -

Unsupervised learning 

Unsupervised learning is a way by which you make the machine learn the hidden factors in the data without providing them any labeled information. So how does it happen? 

Let's take a real-world example, let's say I give the machine 10 images of an apple and 10 images of an orange together mixed up. And by the way, remember the machine doesn't know how an apple looks like or orange looks like, but the machine will be able to cluster the 10 images of an apple together aside and another 10 images of an orange together aside.

How the mission can do that is because it finds similarities between two different images and it will be able to cluster together. 

So the process by which permission learns from the data without being providing any labeled information to this is called Unsupervised learning.  

Unsupervised learning is further subdivided into three categories

 Based on how the model finds hidden patterns within the data

Clustering - Clustering is a concept by machine or model clusters different data voice into groups based on their similarities. For example, customers who like to watch a similar film on Netflix, students who like to play cricket. These clusters can be further analyzed to understand the characteristics of the groups. 

Association -  Association finds out the relationship between two different data points of the data. For example, a customer who has recently bought a house is most likely to purchase the furniture. 

Recommendation system - One of the most widely used and popular, and supervised learning techniques in the market. Netflix and Amazon use customers' past data to recommend to the customer which movie they can watch next.

The third type of Machine learning is - 

Reinforcement learning

Reinforcement learning is making a huge headline these days, reinforcement learning is nothing but making the model to learn through trial and error. Every time the machine or model makes the right decision it gets rewarded and every time the model takes the wrong decision it's get penalized. 

Let's take a real-world example. Think of the mouse inside a maze, the objective of the most is to come out of the mist. But the most can actually take multiple parts to the exit. So how does the mouse finds out what is the right solution? By taking a lot of trials and error the most optimize And finally finds out which is the right direction and as well as the right path to the exit. 

The process by which you make a machine or model to learn through trial and error and by rewarding the model then it takes the right decision and penalizing the model. Then it takes a wrong decision that is reinforcement learning.

Also If You want to know more about What is Artificial Intelligence then Make sure to check out my new post on

Why artificial intelligence is the future? | what is AI?

• FAQ for Machine learning

Q1. What is the difference between machine learning and artificial intelligence?

Artificial intelligence is a technology that mimics natural intelligence. It understands the structure and behavior of data. For example, if it is a photographic image of a person it can identify that it is a person and not something else. But machine learning does not. 

Machine learning is similar but it is designed to mimic and learn the behavior from the data set. The aim is to aid humans in decision-making. A typical example of machine learning is the predictive maintenance of machinery and procedures.

Q2. What is machine learning used for?

Machine learning is used in internet search engines, healthcare, and finance, and more. The answer to the final question is a bit of everything. AI is not just about finding a needle in a haystack, or about being able to translate your dream into a sentence. It's actually a set of technologies, or algorithms, which scientists use to model the way we think.

Q3. Why is machine learning important today?

Machine learning is important because of its wide range of applications and its incredible ability to discover complex patterns even with incomplete data. The data may not be perfect and won't perfectly represent reality, but machine learning can often still find trends or other useful attributes within the data.