Unsupervised Learning in Machine Learning, Artificial Intelligence and Data Mining

Image Source: Read.ai

Unsupervised Learning in Machine Learning, Artificial Intelligence and Data Mining


In this article, we are going to talk about the unsupervised learning, as by its name its define a lot. But, firstly we will talk about Machine Learning, Artificial Intelligence and Data Mining.

Artificial Intelligence:

Artificial Intelligence is the process of replication of human intelligence, as it is clear by its name. Here, we try to make the computer work and act like a human.

Machine Learning:

Coming to Machine Learning, it can be simply described as an application of artificial intelligence (AI) that gives systems the ability to automatically learn and improve from experience without being explicitly programmed.

Data Mining:

Data mining is the process of discovering a pattern and relationship in a large dataset; Data Mining is also known as Knowledge Discovery in Database(KDD).
Now, coming to Unsupervised Learning.

Unsupervised Learning:

Unsupervised learning is finding the hidden pattern from unclassified and unlabelled data. The idea behind this is to make a computer that can identify complex comparisons and patterns in data without much human interference. Unlike the supervised learning, where we provide labelled data(output corresponding to input) to learn and predict.

We call this “unsupervised learning” because we start with unlabeled data(there’s no Y).

READ  Supervised Learning: Classification and Regression Problems

These models do not predict a result, rather focus on the hidden structures, relationships, and interconnectedness of the data.

In unsupervised learning, a model groups messy information according to similarities and differences even though data is non-categorical, in simple language, the data will be clustered/grouped according to the attributes/features of the data points over the given dataset.

For example of unsupervised learning, we will take a bunch of people and then we have to divide them according to there features. Features like the sex of a person, facial hair, height, weight. With the help of any above-given feature, we make some cluster out the people.

The most common unsupervised learning method is cluster analysis. Cluster analysis used for exploring the hidden patterns and grouping the data.

Here in unsupervised learning, there are the different class of Cluster analysis.

We will be focusing on these two famous clustering methods.

  1. k-Means clustering
  2. Hierarchical clustering

k – Means Clustering: 

In K means clustering, groups/clusters are divided into K numbers, where k is defined by the user.


  • Large K numerical value means cluster size will be small.
  • Small K numerical value means cluster size will be larger.

    unsuperviseed learning
    Source: Medium.com

K-means clustering is consist of few steps.

  • Step 1: Set the k centroids
  • Step 2: Find the nearest centroid & update cluster assignments
  • Step 3: Relocate the centroids to the centre of their clusters

Hierarchical Clustering:

Hierarchical clustering is another most used method for cluster formation, as by its name gives us a hint of its work, here in short how this works.

1. Initially, it produces a multilevel hierarchy of clusters by building a cluster tree.

2. Then two nearest clusters are merged into the same cluster.

3. In the end, Hierarchical clustering terminates when there is only a sole cluster left.


Unsupervised learning is quite famous among in all three Machine Learning, Artificial Intelligence and Data Mining as this working as a catalyst for growth and research in these fields. Surely, it has very good and vast future.

Online best courses for the above topics:

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.