In machine learning, building a Anomaly Detection System is important to detect the unusual training set from the examples. Anomaly Detection is the identification of rare item which lies out of the group or from the majority of the data. This lead to the suspicious nature of the training example.
For example:-Suppose we are testing the aircraft by taking down all the aircraft’s data into a graph and one aircraft’s data is found to be located in the different region from the cluster of other aircraft’s data. This example can be considered as the anomaly detection and can be tested further.
Here, the orange dot which corresponds to the training example can be considered as Anomaly Detection.
Application of Anomaly Detection are:-
- Fraud Detection.
- Manufacturing.
- Monitoring computer in data center.
Algorithm of Anomaly Detection:-
The Thesis Whisperer 