Machine Learning by virtue of its name is nothing but training the machines to do some specific task and later helping them learn themselves and keep performing the same task with higher efficiency.

With the advent of the internet, there is so much data generated around the world, this data is of no use for future until we derive an insight from it and use this insight to improve the future process.

Can we derive an insight from data?

Yes. Data has patterns which will help us arrive at a conclusion about the human behavior; as this data is being generated by human activity or on top of human activity.

How these insights help us?

Imagine you have been to a retail shop and bought a bathing soap, toothbrush, body scrub, and a face wash. Now look at your items, did you find any pattern in what you have bought? you have bought items which help you clean your body. As a retail shop owner, I will infer that if I keep other cleanser items near the items you had bought earlier then it is most likely that you will buy other cleanser items too.

So this example is of just one data or one behavior, hence in actual machine learning, we will try to infer meaning from millions of such data sets. Processing and analyzing such huge amount of data is a complex thing and its far from the human capacity.

What is Machine Learning? - Machine Learning Future
What is Machine Learning? – Machine Learning Future

Practical use case of Machine Learning

Suppose, if I ask you, around 1000 buyers have bought a new item under which item’s influence(assuming each buyer had bought on an average of 20 items); what would be your answer? Wait a minute! Did you just think that you will find a common item out of all other 19 items and that should be the influencer? Could be, but if we have no such item which is common among all buyers then? Hence, the question boils down to, how to derive insights from such complex data? Is there any system capable of doing that?

Yes. Thanks to machine learning.

I will keep above question on readers’ table to think how or which machine learning concept can address such problem.