Supervised Learning: A Real -Life Scenario
The Three Musketeers
I hope you got a fundamental understanding of the basic requirements in the previous blog post. To make sure that you don’t get bored of math, here is another cool blog post laying out the primary problems types in the domain of machine learning. However, to keep the blog post short (and keep you awake), I will only focus on one of the three problem types — Supervised Learning. This blog post will also walk you through a real-life scenario so as to engrave the concepts in your brain.
Before that, let me give you a visual depiction of the three divisions.
Okay. Got the names? Good!
Cool, then let’s get right into it.
Supervised Learning
Here’s the defintion:
Supervised learning is the machine learning task of learning a function that maps an input to an output based on example input-output pairs.
I am pretty sure you have heard this term many times if you try to play with the artificial intelligence. In fact, most of the common and easy (some difficult) problems are categorized in the sub-domain of the supervised learning. Some examples include stock market prediction, medical diagnosis, weather prediction, and time-series forecasting.
With the advent of numerous machine learning frameworks incorporated into various programming languages, it has been extremely convenient to generate these type of solutions. Let me give you a pragmatic scenario.
CASE STUDY:
Imagine you are an owner of a company or organization which is flourishing in the market.
Few months down line, you observe a peculiar variation in the number of employees who are resigning from your organization. This causes a massive setback to your business in addition to losing money. You are determined to save the business and thus call in your finest data scientists to help you and your company.
The name of this problem is called Churn Modelling and is an extremely important concept for the retention of employees in an organization.
So, the team of data scientists asks you to gather data on people who are leaving your company and their employment details, such as age, department, salary, marital status, location, work satisfaction level, credit balance, etc.
Then, the team first tries to analyze the features that might affect their work and ultimately, resignation.
After careful analysis, the team of data scientists trains an algorithm(model) which uses the data to find patterns in it.
And voila! The team gives you good news. It was found out that people of certain age groups and a low work satisfaction level are more prone to leave your company and others. In addition, it was also found out that the work environment had a greater effect on the decision.
Understanding this, you decide to make changes to the work environment and incorporate activities to keep the employees happy. And guess what, you just saved yourself millions of dollars, and more, you saved your company from dying.
Though trivial, supervised learning is immensely powerful. But you must remember one thing, data plays a crucial role in case of supervised learning problems and other machine learning problems.
To end the article, let me tell you the primary types of supervised learning problems.
- Regression — The art of predicting continuous values given some input values.
- Classification — The art of segregating entities into distinct groups.
Again, if you have gotten your hands dirty with machine learning you might know these terms. If you haven’t don’t worry, this blog exists for a reason. I’ll explain it all, in a later article of course.
But I hope you enjoyed this article and understood what supervised learning is. Expect unsupervised and reinforcement learning in the next few articles. If you have any doubts, post them in the comments below. Until then,
Upgrade your mind, upgrade your life!
Check out the other article on regression types:
If you like this article, consider buying me a coffee :)
