Most of us have till now heard something or someone talking about "Machine Learning". The next probable question it should bring up is, "What exactly Machine Learning is"?, "Why so popular"?, "What's the buzz all about and why now and not earlier"? All of it makes sense. There are tons of tutorials, webinars and study resources available for Machine Learning these days and they might teach you a thing or two, but the main question is, "Do you know what Machine Learning is"?, "Do you know what you are doing?"
Machine Learning
I'm not going to run you into a Wikipedia copied "technical" definition here, our main focus would be understanding the real meaning and power of Machine Learning. Let's start with "What a machine is?".
According to my favorite Indian movie(3 Idiots) and a common man's understanding, "a machine is anything that reduces human effort!". You can read the technical definitions and more about machines but the core understanding goes to convey the fact that machines are there to reduce human effort in different fields. Now machine learning, as the name itself justifies, is a technique through which machines "learn". Now, what is it that they learn, and what do we as humans have at the end of this process?
Machine Learning is equipped with a whole bunch of algorithms and they are constantly improving to give better results! Machine Learning is a subset of Artificial Intelligence and it is used to achieve artificial intelligence by following a certain process of operations. In the simplest terms, machines recognize the patterns from the data which is fed to them in a certain format (which requires data preprocessing) and start to learn these patterns in an attempt to predict a certain label or result (or a computation in simpler terms) and finally when we think that our machine learning model is good enough and it starts making right predictions on the test and custom data, we can put the model in the process of deployment in the real world.
For instance, suppose a team of doctors in a certain hospital wants to use Machine Learning to predict whether a person has heart disease or not given some crucial parameters regarding their health. The Machine Learning Engineer or the Data Scientist who has been assigned the task to make this model would need some data from the hospital staff along with the result label. In simpler terms, he/she would need the data regarding all the parameters which affect a human's heart health and whether a patient with a given set of parameters has heart disease or not. He/she would preprocess the data and pass it to the Machine Learning algorithm which is best suited for this use case and after training the model, he/she would test the model's accuracy of predictions. After a certain accuracy has been reached and it's confirmed that the model is actually predicting reliable information, then only it can be deployed for real-world use. This is an example of a "Classification" problem.
I highly recommend checking this Youtube video for a visual understanding of Machine Learning.
The problems addressing the concept of Machine Learning can further be divided into categories and that is something I'm going to talk about in the next post.
Hope you enjoyed reading this!