Classification And Novel Class Detection Approaches Of Feature Evolving Data Stream

1716 Words Nov 25th, 2014 7 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 stream data classification processes and method of these classification techniques. Different authors used different method such as data miner and tree based approach for reduced such types of issues. Ensemble of classifier is used to detect novel classes for feature evolving data stream.

Keywords: data stream, novel class, classification technique, outlier.

INTRODUCTION

Data stream classification is more difficult due to its dynamic changing nature as compared to stationary data. First data streams are in infinite length so it is not feasible to use all the historical data for training. For this issue multi-pass learning algorithms are not applicable. Incremental learning approach is well suited for this problem. Second concept drift,…
Open Document