Classification And Novel Class Detection Approaches Of Feature Evolving Data Stream

1716 Words7 Pages
A Survey On Various Classification And Novel Class Detection Approaches Of Feature Evolving Data Stream Abstract: The classification of data stream is challenging task for data mining community. Dynamic changing nature of data stream has some difficulties such as feature evolution, concept evolution, concept drift and infinite length. As we know that the data streams are huge in amount, it is impractical to store and use all the data for training. Concept drift occurs when underlying concept changes. Concept-evolution occurs as a result of new classes evolving in the stream. Another important characteristic of data streams, namely, feature evolution, in data stream new features emerge as stream advancement. In this paper we discuss the…show more content…
Various techniques have been proposed to address this problem. In order to deal with concept drift, classification model must be updated with recent data. Another characteristic of data stream is concept evolution, when new classes evolve in data concept evolution occurs. In order to deal with this problem classification model must be able to detect novel classes when they appear. For example intrusion detection in a network traffic. Most important characteristic is feature evolution in which new features (words) emerge and old features fade away. Ensemble techniques have been more popular than single model [1]. In this technique more than one classifier is used for classification with higher efficiency. Each classifier in the classification model is trained on different data chunks. With the help of advanced data streaming technologies [2], we are now able to collect large volume of data for different application domains. For example credit card transaction, network traffic monitoring etc. the presence of irrelevant and redundant data slows down the learning algorithms [3] [4]. By removing or ignoring irrelevant and redundant feature, prediction performance and computational efficiency can be improved. Multiclass miner works with dynamic feature vector and detects novel classes. It is a combination of OLINDDA and FAE approach. OLINDDA and FAE are used to detect novel classes and to classify data chunks
Open Document