Little Bits of Artificial Intelligence

1. What is K-Nearest Neighbors (KNN) Regression?

K-Nearest Neighbors Regression is a simple machine learning algorithm that predicts a value (like a house price or temperature) based on the values of the K most similar (nearest) known data points.
It doesn’t build a formula or model — it just looks around at the most similar examples and takes the average of their values to make a prediction.

2. Real-Life Story

1: House Price Estimator – The Friendly Realtor

Imagine this: A realtor wants to sell his 2-bedroom house in a new town, but he doesn’t know how much to ask.

If we go to a local real estate advisor. He doesn’t use complex formulas. Instead, he says:

“Let me check the 3 most similar houses that were sold recently — same size, similar neighborhood, similar features.”

He finds 3 houses nearby:

• One sold for ₹55 lakhs
• One for ₹60 lakhs
• One for ₹58 lakhs

He averages them: (55 + 60 + 58) / 3 = ₹57.67 lakhs

So, he tells us: “You should price your house around ₹57.67 lakhs.”

That’s KNN Regression with K = 3. It didn’t use a formula — just nearby examples!

2: Health App – Predicting Blood Sugar

Imagine: A health app wants to predict a person’s blood sugar level based on their:

• age
• weight
• breakfast type
• last night’s sleep hours

Now, a new user opens the app and logs their data:

• Age: 45
• Weight: 72 kg
• Ate oats for breakfast
• Slept 6 hours

The app checks the 5 most similar users in its database (based on age, weight, etc.), and finds their sugar levels:

• 98, 102, 97, 100, and 103 mg/dL

It averages them: (98 + 102 + 97 + 100 + 103) / 5 = 100 mg/dL

So the app predicts: “Your estimated blood sugar is 100 mg/dL.”

Again, this is KNN Regression with K=5.