Understanding K-Nearest Neighbors (KNN):My Beginner-Friendly Explanation

Today I learned how the K-Nearest Neighbors (KNN) algorithm works, and this is my simple explanation of it from a beginner’s perspective. KNN is one of the easiest machine learning algorithms to understand because it makes predictions purely based on distance — basically, “who is closest to the test point.”

Let me walk you through everything I learned step by step.

What Is KNN?

KNN (K-Nearest Neighbors) is a supervised machine learning algorithm used mainly for:

• Classification (predicting a category)
• Regression (predicting numbers)

The idea is simple:

👉 When you give a new data point, KNN looks at the K closest points in the training set. 👉 It checks their labels. 👉 The most common label becomes the prediction.

No training phase. No model training. Just distance + neighbors + voting.

That’s why KNN is sometimes called a lazy learning algorithm.

Step 1: Importing Required Libraries

IMPORTING LIBRARY FILES

• NumPy → for numerical calculations
• Counter → to count votes from neighbors

This is all I needed to build KNN .

Step 2: Creating the Distance Function

This function calculates the Euclidean Distance (straight-line distance). I learned that “nearest” simply means:

✔ The smallest distance ✔ The closest point

Step 3: Writing the KNN Prediction Function

Here’s what happens:

✔ Find distance: Loop through all training points and calculate how far each one is from the test point.

✔ Sort : Sort all points so the closest ones come first.

✔ Select top K neighbors : Pick the first K labels — these belong to the nearest neighbors.

✔ Voting: Whichever label appears the most wins.

This gives the final prediction.

Step 4: Preparing Data

Here:

• The first three points belong to class A
• The last two belong to class B
• I want to predict the label for [4,5]

Step 5: Final Prediction

OUTPUT: A

The three nearest neighbors of [4,5] all belong to Class A, so KNN predicts A.

This is exactly how KNN works.

Conclusion :

Learning KNN gave me a clear idea of how distance-based machine learning works. The algorithm is incredibly easy to understand:

1. Pick a value for K
2. Find the closest K neighbors
3. Let them vote
4. That becomes the prediction