Multiclass classification using support vector machine and OligoIS Technique

International Journal of Computer Trends and Technology (IJCTT)          
© 2015 by IJCTT Journal
Volume-21 Number-1
Year of Publication : 2015
Authors : Shubhangi Dusane, Er.Praveen Bhanodia, Er.Atish Mishra


Shubhangi Dusane, Er.Praveen Bhanodia, Er.Atish Mishra "Multiclass classification using support vector machine and OligoIS Technique". International Journal of Computer Trends and Technology (IJCTT) V21(1):14-18, March 2015. ISSN:2231-2803. Published by Seventh Sense Research Group.

Abstract -
In current era, data were collected from various sources like, public survey, web survey, data collected by analysis etc. So these collected data may be or may not be having balanced class. That time, this is typical, task to manage data. If, we want select some instances from these data, then this is a triodes job. Here proposed Hybrid approach for the large class problem in multiclass classification using SVM for solving large class problem. Hybrid OligoIS is an extension of OligoIS. Proposing hybrid approach of OligoIS and SVM can be used in large class problem and also it better perform when increases size of the database. On the Er.Praveen Bhanodia basis of Binary Decision Tree and Probabilistic output of Support Vector Machine here want to present Binary Decision Tree SVM (BDT) using Support Vector Machine(SVM) as an original approach to the multi-class classification problem. Instead of using a simple SVM classifier in each node, here propose SVM classifier associated with a threshold function of OligoIS to estimate the probability of membership to each sub-group in the node. Proposing 6 classifier architecture OligoIS-SVM takes the advantage of both the highest classification accuracy of SVM and the efficient computation of the tree architecture. OligoIS-SVM is based on recursively dividing the classes into two groups in every node of the Binary Decision Tree and training an SVM associate with threshold value (OligoIS Algorithm). Here we take three types of data set small, medium and large. There are three Dataset in small dataset category i.e. IRIS, WINE, ECOLI number of classes less than 10. and performed our algorithms on those dataset, also in medium dataset number of classes less than 50. We have taken dataset i.e. Movement Libra and finally in large number of dataset we have taken two dataset Thyroid, Yeast. Number of classes more than 100 and performed our functionality process.

[1] Agarwal, S., Pandey, G. N., & Tiwari, M. D.:Data Mining in Education: Data Classification and Decision Tree Approach.(2012)
[2]. Nicolas Garcia-Pedrajas, Javier Perez-Rodriguez, Aida de Haro-Garcia , ―OligoIS: Scalable Instance Selection for Class-Imbalanced Data Sets‖, IEEE Transactions On Systems, Man, And Cybernetics—Part B: Cybernetics
[3].Dr.D.Ramyachitra, nikandan,‖Imbalanced DataSet Classification and solution:A Reviw‖ International Journal of Computing and Business Research (IJCBR) ISSN (Online) : 2229-6166 Volume 5 Issue 4 July 2014 [4]. Anyanwu, M. N., & Shiva, S. G.: Comparative analysis of serial decision tree classification algorithms. Vol.3, 230-240 International Journal of Computer Science and Security(2009)
[5]. Asuncion, A., & Newman, D. J. UCI Machine Learning Repository. Irvine, CA: University of California. School of Information and Computer Science. 2007
[6]. Bakar, A. A., Othman, Z. A., & Shuib, N. L. M. : Building a new taxonomy for data discretization techniques. In Data Mining and Optimization, 2nd Conference. 132-140. IEEE ( 2009)
[7]. Balagatabi, Z. N., & Balagatabi, H. N.: Comparison of Decision Tree and SVM Methods in Classification of Researcher's Cognitive Styles in Academic Environment. vol.1, 31-43. Indian Journal of Automation and Artificial Intelligence, (2013)
[8]. Bramer, M.: Decision Tree Induction: Using Entropy for Attribute Selection. In Principles of Data Mining. 49-62. Springer London. (2013)
[9]. Bunkar, K., Singh, U. K., Pandya, B., & Bunkar, R.. : Data mining: Prediction for performance improvement of graduate students using classification. In Wireless and Optical Communications Networks (WOCN), Ninth International Conference, 1-5. IEEE. (2012)
[10]. Burrows, W. R., Benjamin, M., Beauchamp, S., Lord, E. R., McCollor, D., & Thomson, B.: CART decision-tree statistical analysis and prediction of summer season maximum surface ozone for the Vancouver, Montreal, and Atlantic regions of Canada. Vol. 34 1848-1862, Journal of applied meteorology( 1995)
[11]. Cortes, C., & Vapnik, V.: Support vector machine. Vol.203, 273-297. Machine learning (1995).49
[12]. Cover, T., & Hart, P.: Nearest neighbor pattern classification. Information Theory. Vol. 13, 21- IEEE Transactions (1967)
[13]. Dasarathy, B.V.: Nosing Around the Neighborhood: A New System Structure and Classification Rule for Recognition in Partially Expos Environments. Vol. PAMI-2, No. 1, 67-71. Pattern Analysis and Machine Intelligence. IEEE Transactions (1980)
[14]. de Oña, J., López, G., & Abellán, J.: Extracting decision rules from police accident reports through decision trees.Accident Analysis & Prevention. Elsevier (2012)
[15]. Devijver, P. A., & Kittler, J.: Pattern recognition: A statistical approach. Englewood Cliffs, NJ: Prentice/Hall International.(1982)
[16]. Everitt, B. S., Landau, S., Leese, M. and Stahl, D.: Miscellaneous Clustering Methods, in Cluster Analysis, 5th Edit., JohWiley and Sons, Ltd, Chichester, UK. (knn).