What is Machine Learning?

What is Machine Learning ?

This question comes in our mind when we read or hear this word many times in a day. But why 75% of the world talking about Machine Learning and why there is tons of Job Post in Machine Learning.

Machine Learning Introduction

Let’s understand What is Machine Learning in a very simple and practical way. Probably we use Machine Learning many times in a day but we don’t know actually.

Every time we search something on Google Search Engine and it gives us some accurate result also a part of Machine Learning.

Facebook recognizes our friends on Photos and sends us friend notification also a part of Machine Learning.

Gmail send our spam email in the spam folder and we even don’t care about those emails. Gmail learnt over the time by our email checking behaviour and knows very well about spam emails. It’s all part of Machine Learning.

Definition

Tom Mitchell provides a more modern definition: “A computer program is said to learn from experience E with respect to some class of tasks T and performance measure P, if its performance at tasks in T, as measured by P, improves with experience E.”

Generally, Machine Learning is the process that provides learning ability to computers and improves from past experiences for accurate predictions. Machine Learning basically a computer program that assesses the data and learns from them for better predictions.

The learning prediction could be to recognize friends in Facebook, recognize the animal, recognize fruits like a banana or apple in a photo.

Basically Machine Learning program or algorithm uses statistics to find a pattern in a massive amount of data. And data could be anything like text, images, number etc and this data fed to process into Machine Learning program.

The basic aim is to allow computers to learn themselves by data and do actions based on data learning.

For example, Youtube, Netflix and other video platforms do behaviour analysis of user data and recommended us videos based on data.

Types of Machine Learning

Machine Learning algorithm categorizes in three types supervised, unsupervised and reinforcement learning.

Supervised Machine Learning

In supervised Machine Learning, the program is exposed to labelled data like animals images with their names so that system can recognize a cat, dog and another animal by their photos.

Labelled data actually tell the machine to look for an exact pattern and provide result based on that pattern.

Supervised learning takes a huge amount of differently labelled data to study the pattern. Google’s Image and youtube videos are vast labelled dataset and they are using it for supervised learning for a more accurate result.

Unsupervised Machine Learning

Like supervised learning, there is no labelled data to look for any pattern. The machine just looks for pattern whatever it can find on similarity and then categorized them in smaller groups.

Google News is a good example of unsupervised learning that put together similar news feed daily.

Reinforcement Machine Learning

Reinforcement Machine Learning is a method that works on trial and error to achieve a clear goal. This method allows machines and programs to automatically determine the best behaviour within a specific context in order to maximize its performance.

This learning works on reward feedback method that is required for the program to take feedback on error and learn.

Machine Learning Examples

Database Mining is the greatest example of machine learning. Companies do automation on larger dataset like web click data analysis, medical records, Netflix, Amazon, computer vision and NPL.

In web click data, machine learning is used to identify the user behaviour what they look for, and what type of results a user clicks so by this machine learning algorithm, companies determining the user behaviour on the web.

Medical records keep tons of user medical data set of medical reports. A machine learning program turns those medical reports in raw data and does analysis the disease symptoms to study further.

Netflix & Amazon is also using machine Learning for product recommendation. They use machine learning for user behaviour on their past movies and past products purchases.

Natural language processing and computer vision also part of machine learning. That we will discuss in the coming chapters of machine learning

Machine Learning is a vast technology to learn. I will learn more and share more in the coming days with a simple approach and examples.