Unsupervised Learning in AI

Unsupervised Learning in AI

1. Unsupervised Learning
   1. Introduction
   1. Clustering
   1. Comparison between Supervised, Unsupervised, and Reinforcement Learning.

Unsupervised Learning

This is the third major category of Machine Learning. Unsupervised learning happens when we have data without additional feedback, and patterns.

For example, in a case of supervised learning, we always had the labeled data, like whether a data point represents rain or No rain, and using those labels, we were able to define the relationships between the data and make further predictions.

In unsupervised learning, we do not have any labels, for example we are given the data of temperature and humidity of 30 days, but that does not have label if it rained or not on that particular day, these kind of problems falls under the category of unsupervised learning.

In simple words, in unsupervised learning problem we do not have output labels but still we want to define some relationships, some patterns in the data and that can be done by Clustering.

Clustering is the task of organizing the set of objects into distinct clusters or say group of objects in such a way that all the similar objects tend to be in the same group or cluster.

You can learn more clustering here.

And, implementation of Clustering Algorithm here.

Supervised Learning	Unsupervised Learning	Reinforcement Learning
Machine Learning model is trained on labeled data.Types of problems solved by supervised learning are Regression and Classification.Training the supervised Machine learning model requires external supervision.	Machine Learning models are trained on unlabeled data.Types of problems solved by unsupervised learning are Clustering and Association.No supervision is required for the training of the unsupervised machine learning model.	AI agent start taking random actions and then learn from errors and rewards.Types of problems solved by reinforcement learning is reward-based.No supervision is required for the training the reinforcement machine learning model.

Comparison between Supervised Learning, Unsupervised Learning, and Reinforcement Learning.

Comparison between Supervised, Unsupervised, and Reinforcement Learning on the basis of some examples.

	Supervised Learning	Unsupervised Learning	Reinforcement Learning
Banking	To predict whether the banknote is counterfeit or not.	Segment customers by behavioral categories.	To create the next best offer model for the call center group.
HealthCare	To predict whether is diabetic or not.	To categorize the MRI data by normal and abnormal images.	To allocate scarce medical resources to handle different ER cases.
Retail	To analyze the products which customers buy together?	To recommend products to customers based on their past purchases.	To reduce excess stock with dynamic pricing.