Unsupervised Learning is a machine learning technique where not all variables and data patterns are classified. Unsupervised Learning can be thought of as self-learning where the algorithm can find previously unknown patterns in datasets that don’t have any sort of labels. It is also called as self-organization allows for modelling of probability densities over inputs.
Unsupervised learning is different from Supervised learning
| PARAMETERS | UNSUPERVISED LEARNING | SUPERVISED LEARNING |
| DATASETS | UNLABELLED DATA | LABELLED DATA |
| METHOD OF LEARNING | ALGORITHM LEARN BY ITSELF USING DATASETS | GUIDED LEARNING |
| COMPLEXITY | COMPUTATIONAL COMPLEX | SIMPLER METHOD |
| ACCURACY | LESS ACCURATE | MORE ACCURATE |
Types of Unsupervised learning
1. Clustering
Clustering is a type of unsupervise learning, which makes the groups of objects in such a manner that the objects of same features or pattern (it may be the colour, shape, size etc.)are in one group i.e., the objects of the same group are similar to each other than the objects in other groups. It creates clusters.
For Example
If you want to divide the students but you don’t know how to create these groups. So, clustering can help us to create a different students group based on their marks, creativity, etc.
Some algorithm of Clustering
- K-means
- Hierarchical
- EwKm and many more.
2. Association
Association is a kind of unsupervise learning where you can find the dependencies of one data item to another data item and map them such that they can help you profit better.
For Example
Using a database of transaction in a supermarket to find the items that are brought together eg., buys bread and butter usually buys milk. This could develop a market layout
Some algorithms of the association are
- Apriori
- Fp growth
Application of unsupervised learning
- Recognize patterns to cluster data
- Defects in data collected
- Identify dependencies
- Cleansing the datasets by removing unwanted features
