Title: Performance analysis of Hoeffding trees in data streams by using massive online analysis framework
Authors: P.K. Srimani; Malini M. Patil
Addresses: R & D Division, Bangalore University Jnana Bharathi, Mysore Road, Bangalore-560056, Karnataka, India ' Department of Information Science and Engineering, J.S.S. Academy of Technical Education, Uttaralli-Kengeri Main Road, Mylasandra, Bangalore-560060, Karnataka, India
Abstract: Present work is mainly concerned with the understanding of the problem of classification from the data stream perspective on evolving streams using massive online analysis framework with regard to different Hoeffding trees. Advancement of the technology both in the area of hardware and software has led to the rapid storage of data in huge volumes. Such data is referred to as a data stream. Traditional data mining methods are not capable of handling data streams because of the ubiquitous nature of data streams. The challenging task is how to store, analyse and visualise such large volumes of data. Massive data mining is a solution for these challenges. In the present analysis five different Hoeffding trees are used on the available eight dataset generators of massive online analysis framework and the results predict that stagger generator happens to be the best performer for different classifiers.
Keywords: data mining; data streams; static streams; evolving streams; Hoeffding trees; classification; supervised learning; massive online analysis; MOA framework; massive data mining; MDM; dataset generators; performance evaluation.
International Journal of Data Mining, Modelling and Management, 2015 Vol.7 No.4, pp.293 - 313
Published online: 27 Dec 2015 *Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article