Unsupervised Learning

 Unsupervised Learning:

In this Unsupervised learning we do not have any target or output variable, it is completely an unlabeled data. We do not know any prior knowledge about what our results should look like.

Types of Unsupervised Learning:

1. Clustering
2. Visualization Algorithms
3. Anomaly Detection

Clustering:

The most commonly used unsupervised learning method is cluster analysis. It is used to find data clusters so that which are more similar characteristics.

For instance, say you have lot of data about your blog's visitors.You may want to learn a clustering algorithm to try to detect group of similar visitors. Let's say you don't tell the algorithm which group a visitor belongs to: it finds those connections with out your help. It helps you to target customers. For example you might notice that

• 40% of  your visitors are male who love comic books and generally read your blog in the evening.
• While 20% are young Sci-fi lovers visit your blog during in the weekends, and so on.

Examples:

Online news portals segments articles into different categories like Political, Sports, Business and etc..

Visualization Algorithms:

Visualization algorithms are unsupervised learning algorithm  that accepts unlabeled data and display this data in intuitive 2 D  or  3 D format. The data is separated into some what clear to understand. In the figure, animals are rather well separated from vehicles and Horses are close to deer but far from birds and so on. 

Anomaly Detection:

This algorithm detects anomalies in data with out any prior training. It can detect suspicious credit card transactions and differentiate a criminal from a set of people.