Scribd
Upload
43 views4 pages

Understanding Unsupervised Learning Techniques

This document provides an overview of machine learning, specifically focusing on unsupervised learning, which involves learning from unlabeled data to identify patterns. It explains various types of machine learning, including supervised, unsupervised, and reinforcement learning, along with real-life applications and common techniques such as clustering, dimensionality reduction, and association rule learning. Additionally, it details popular clustering algorithms like K-Means and Hierarchical Clustering, outlining their processes and objectives.

Uploaded by

22cseaiml096.potnuruvamsikrishna
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd
43 views4 pages

Understanding Unsupervised Learning Techniques

This document provides an overview of machine learning, specifically focusing on unsupervised learning, which involves learning from unlabeled data to identify patterns. It explains various types of machine learning, including supervised, unsupervised, and reinforcement learning, along with real-life applications and common techniques such as clustering, dimensionality reduction, and association rule learning. Additionally, it details popular clustering algorithms like K-Means and Hierarchical Clustering, outlining their processes and objectives.

Uploaded by

22cseaiml096.potnuruvamsikrishna
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as DOCX, PDF, TXT or read online on Scribd

UNSUPERVISED LEARNING

 What is Machine Learning?


 Machine Learning (ML) is a branch of Artificial
Intelligence (AI) that focuses on creating systems that
can learn from data and improve their performance
over time without being explicitly programmed.

 Machine learning is a method by which computers learn


from experience (data) to make predictions or decisions
without being directly told how to do it.

 How it works?
 Instead of writing code to tell the computer what to do
step-by-step, we give it data and a model, and it
figures out the patterns or rules on its own.
 Types of Machine Learning:
 Supervised Learning – Learn from labeled data.
Example: Predicting house prices from past data
(features + price).
 Unsupervised Learning – Find patterns in unlabeled
data.
Example: Customer segmentation in marketing.
 Reinforcement Learning – Learn by trial and error with
rewards.
Example: A robot learning to walk.
 Real-Life Applications:
 Spam email detection
 Movie or product recommendations
 Facial recognition
 Self-driving cars
 Fraud detection in banking
 What is Unsupervised Learning?
 Unsupervised Learning is a type of machine learning
where the model learns from data that is not labeled —
meaning there are no predefined categories or
outcomes.
 In supervised learning, the data comes with answers
(like "spam" or "not spam").
In unsupervised learning, there are no answers — the
algorithm tries to find hidden patterns or structures in
the data on its own.

Example:
Imagine you have a bunch of customer data (age,
income, purchase habits), but you don’t know anything
about who they are.
With unsupervised learning, the algorithm might:
 Group similar customers together (this is called
clustering).

 Find unusual behavior (called anomaly detection).

 Common Techniques in Unsupervised Learning:

 Clustering – Group similar data points.

Example: Customer segmentation, grouping users by


behavior.

Algorithms: K-Means, Hierarchical Clustering.

 Dimensionality Reduction – Simplify data without losing


important information.

Example: Visualizing high-dimensional data in 2D.

Algorithms: PCA (Principal Component Analysis), t-SNE.

 Association Rule Learning – Discover relationships


between features.

Example: Market basket analysis (people who buy


bread often buy butter).

Algorithm: Apriori

 What is Clustering in Machine Learning?


 Clustering is an unsupervised learning technique that
involves grouping similar data points together based on
their features — without using any labels.

Simple Explanation: Imagine you have a basket of


mixed fruits, but they’re not labeled.
Clustering is like automatically grouping the fruits into
"apples", "oranges", and "bananas" based on their
shape, size, and color — without knowing the names.

 Objective of Clustering:

To divide a dataset into clusters, where:

 Data points in the same cluster are more similar to each


other.
 Data points in different clusters are more different.

 Popular Clustering Algorithms:

 K-Means Clustering

1. You choose K (number of clusters).


2. The algorithm tries to find K groups by
minimizing the distance within clusters.
3. Fast and widely used.

 Hierarchical Clustering

1. Creates a tree-like structure of clusters.


2. No need to choose K upfront.
3. Good for visualizing with dendrograms.

 DBSCAN (Density-Based Spatial Clustering)

1. Groups based on density of points.


2. Can find clusters of different shapes and sizes.
3. Good at finding outliers.

 K-Means Clustering

K-Means Clustering is an unsupervised machine learning


algorithm used to group similar data points into K distinct
clusters.

How K-Means Works (Step-by-Step):

1. Choose the number of clusters (K): You must decide


how many groups you want to divide the data into.
2. Initialize cluster centroids: Randomly pick K points from
the dataset as the initial centroids (cluster centers).
3. Assign each data point to the nearest centroid: Use
Euclidean distance to measure closeness.
4. Recalculate centroids: For each cluster, compute the
new centroid as the mean of all points in that cluster.
5. Repeat steps 3 & 4: Until the assignments don’t change
anymore (convergence), or a max number of iterations
is reached.

Mathematics Behind It:

 Distance formula:

For a point x=(x1,x2) and centroid c=(c1,c2):

Distance=square root of (x1−c1)2+(x2−c2)2

 Hierarchical Clustering

Hierarchical Clustering is an unsupervised learning method


that builds a tree-like structure (called a dendrogram) to
group data points based on similarity.

There are two main types:

1. Agglomerative (Bottom-Up) – Most Common

 Start with each data point as its own cluster.


 Merge the two closest clusters.
 Repeat until all points belong to a single cluster.

2. Divisive (Top-Down)

 Start with one big cluster.


 Split it recursively into smaller clusters.
 Stop when each point is in its own cluster.

Common questions

Powered by AI

Anomaly detection in unsupervised learning involves identifying outliers within data that do not conform to expected patterns, which can be crucial for detecting fraudulent activities in finance or identifying security breaches in cybersecurity. By learning the normal behavior of systems and flagging deviations, organizations can proactively address potential threats and improve operational efficiency and safety .

Hierarchical Clustering provides a way to visualize the clustering process through dendrograms, illustrating the data structure and relationships, unlike K-Means which outputs fixed number clusters. It doesn't require specifying the number of clusters upfront and can capture hierarchical relationships, which is beneficial for understanding nested groups within the data .

The main objectives of clustering in machine learning are to group similar data points, ensuring intra-cluster similarity and inter-cluster dissimilarity. These objectives facilitate the understanding of data structures, allowing for efficient data analysis and informed decision-making by identifying patterns, trends, and anomalies within datasets that can lead to strategic business insights .

Unsupervised learning techniques can struggle with noisy or complex datasets due to their reliance on finding hidden patterns without labeled guidance. Noise can obscure actual data structures, leading to inaccurate clustering or misleading insights. Complex datasets with variable densities or irregular distributions might result in inefficient clustering or overfitting, emphasizing the need for preprocessing and suitable algorithm selection .

K-Means Clustering works by dividing a dataset into K distinct clusters. Initially, K points are selected as centroids, then data points are assigned to the nearest centroid based on the Euclidean distance. Centroids are recalculated as the mean of assigned points, and the process repeats until convergence. Advantages of K-Means include its speed and simplicity, but it requires specifying the number of clusters upfront and may not handle non-linear boundaries or clusters of varied shapes well .

Unsupervised learning differs from supervised learning as it works with unlabeled data, meaning there are no predefined outputs or categories for training the model. In contrast, supervised learning utilizes labeled data to learn the relationship between input and output. Real-world applications of unsupervised learning include customer segmentation and anomaly detection, while supervised learning is used in scenarios like predicting house prices or spam detection .

Density plays a crucial role in DBSCAN clustering, as it groups points based on densely packed regions, making it effective for identifying clusters of varying shapes and sizes. It can automatically discover arbitrary-shaped clusters and find outliers, unlike K-Means which is limited to spherical clusters and needs a predefined number of clusters, making DBSCAN preferred for datasets with irregular or complex geometries .

Common techniques in unsupervised learning include clustering, dimensionality reduction, and association rule learning. Clustering groups similar data points, such as customer segmentation, using algorithms like K-Means and Hierarchical Clustering. Dimensionality reduction simplifies complex data into lower dimensions without losing significant information, using methods like PCA or t-SNE. Association rule learning finds relationships between features, such as identifying items frequently bought together in market basket analysis, using algorithms like Apriori .

Unsupervised learning improves over time by identifying and learning from patterns and structures within unlabeled data. Instead of following explicit programming instructions, algorithms continuously adjust parameters based on feedback mechanisms or detected variations, thereby enhancing the model's ability to understand complex relationships and adapt to unseen data .

Dimensionality reduction processes simplify complex datasets by reducing their number of attributes while retaining important structure and variance. This is significant for data visualization, as it allows for easier interpretation of high-dimensional data in 2D or 3D forms. Common algorithms used for dimensionality reduction include Principal Component Analysis (PCA) and t-distributed Stochastic Neighbor Embedding (t-SNE).

You might also like