Unsupervised ML — AAMIR P

2 min readJul 10, 2023

Hey all!

In this article, we will discuss UnSupervised ML. Unsupervised ML is a branch of AI that trains itself by recognizing patterns and structures in data without the help of labelled examples. It first explores and identifies inherent patterns or relationships and then groups them accordingly.

The above picture is demonstrating two groups of fruits bananas and pineapples. The model is learning its structure and patterns and then groups them accordingly. That’s why, bananas are separate, and pineapples are separate in output. This is a kind of clustering that happens here.

Let’s discuss a real-time example to understand it better and learn process by process. Imagine that you are a manager of an e-commerce platform. You have all customer data from purchase, age, date, etc.

What will you do with this data? You must gain insights from this data to improve your business.

Now, I use unsupervised ML to perform a customer segmentation analysis. Based on similarities in data or customer behaviour the model groups them and exhibits them as clusters. With this data, you can personalise your product, improve marketing, and optimise the inventory.

So, what’s the process the model undergoes?

The data is cleaned, preprocessed and then transformed according to the required format. Then we use k-means clustering or hierarchical clustering to uncover hidden patterns within the customer data.

The algorithms analyze the data and group customers based on their purchasing patterns, preferences, or other relevant features. Each cluster represents a distinct segment of customers with similar behaviours or characteristics. Through this analysis, you may discover different customer segments, such as “frequent buyers,” “budget shoppers,” “luxury seekers,” or “tech enthusiasts.” Each segment may have its own unique set of preferences and requirements.

To summarise the topic,