Unsupervised learning is a type of Artificial intelligence (AI) using information that is neither labeled nor structured and not classified. Allowing the algorithm to act on the information without guidance.
In unsupervised learning, an AI system act with uncategorized unlabelled and unstructured data, an AI system act on data without prior training. Output comes out dependent on the coded algorithms. It is a way testing of AI.
No trainer is provided to an AI system there for the system is restricted to find a hidden structure, similarities differences, etc. unsupervised learning more unpredictable, it can perform more complex processing tasks than supervised learning.
It is like an adult man, no longer to guide him at every step. He learns from his observation that unsupervised learning works. If you have to deals with a lot of data and you don’t of references then unsupervised learning
Unsupervised learning Algorithm list includes:
- K- Means
- Apriori algorithm
- Singular value decomposition
- Independent component analysis
Unsupervised learning example:
Let’s discuss the example of clustering
NASA discovers new astronomical bodies where they meet totally stranger, now we can classify them into two groups previously found an object and new object using unsupervised learning.
Suppose, In a crowd of blonde people and black people. Now given data is crowd people AI system act on given data, it will discover the different group of blonde people and black people. This classification can be based on color.
Example of Association rule is referred to a relation between mobile and mobile cover. The example of mobile and mobile cover seems like fiction, men go shopping to buy a mobile are also likely to buy a mobile cover. A shop has 500 customers per day about 400 customers are buying with cover and 100 customers are only buying a mobile without cover. AI system gives the result that 90% of mobile purchases including a mobile cover.
- Unsupervised learning is a machine learning technique, where no need to provide labeled input data to the system.
- The best application of unsupervised learning is to identified hidden patterns.
- Example of unsupervised learning is the following: 1) Clustering 2) Association 3) K-Means
- The Association rule helps to found complicated associations from large databases.
- The drawback of unsupervised learning is that the system cannot give precise information regarding input data.