Machine Learning: Supervised vs. Unsupervised Learning|Dileenet A/L ICT

In our previous post, we discussed what Machine Learning is and noted that it can be divided into three main types:

• Supervised Learning
• Unsupervised Learning
• Reinforcement Learning

In today’s post, we will focus on Supervised Learning and Unsupervised Learning.

Supervised Learning

As we mentioned before, Machine Learning is about teaching a machine using past data to help it make decisions in the future. In Supervised Learning, we use Labeled Data.

We train a Machine Learning Model. Think of a "Model" as a file where the machine stores everything it has learned during the training process.

Let’s look at an example:

Suppose we want to train a model to determine whether a student has Passed or Failed based on their marks. Look at the table below, which shows the marks for different subjects and the resulting outcome:

Name	Sub1	Sub2	Sub3	Result
James	80	60	65	Pass
Peter	70	54	63	Pass
Ann	61	47	86	Fail
Kate	50	50	64	Pass

In this table, we have labeled data. This means we have the Inputs (marks for three subjects) and the corresponding Output (the result: Pass or Fail).

First, we must train the machine learning model using this data. The machine identifies a Pattern to understand what combination of marks leads to a "Pass" and what leads to a "Fail." The machine stores this pattern in the model. Once trained, we can give the model new marks, and it will predict whether that student passed or failed.

Sometimes, these predictions might be wrong. To increase accuracy, we need to train the model with a much larger volume of data.

Common Supervised Learning Algorithms:

• Linear Regression
• Logistic Regression
• Support Vector Machine (SVM)
• K-Nearest Neighbors (KNN)
• Naïve Bayes
• Decision Tree
• Random Forest

 

Unsupervised Learning

In Unsupervised Learning, we do not use labeled data to train the machine. While we provide data to the machine, the outputs (labels) are not defined.

Let’s look at an example:

If we were using Supervised Learning, we would tell the machine "this is a square" and "this is a circle." However, in Unsupervised Learning, since there are no labels, the machine analyzes the dataset to find its own patterns. It categorizes the data based on what it observes, even though it doesn't "know" the names of the shapes.

The machine might group the data by Shape (circles vs. squares) or by Color.

In Unsupervised Learning, this process of grouping data based on similarities is called Clustering.

                                        Cluster 1: (e.g., all circles)      Cluster 2: (e.g., all squares)

Common Unsupervised Learning Algorithms:

• K-Means Clustering
• Hierarchical Clustering
• Principal Component Analysis (PCA)
• Singular Value Decomposition (SVD)
• Independent Component Analysis (ICA)

 

In this post, we covered the basics of Supervised and Unsupervised learning. I hope you now understand the key differences between the two.

See you in the next post!

Previous post link

https://e-learnict.blogspot.com/2026/01/dogs-cats-and-data-understanding.html

• #SupervisedLearning
• #UnsupervisedLearning
• #Clustering
• #Classification
• #LabeledData
• #PatternRecognition
• #Algorithms