Belief-based chaotic algorithm for support vector data description
- 80 Downloads
One of the efficient tools to handle segregation of imbalanced data is support vector data description (SVDD). In contrast to support vector machine (SVM), enclosing target data in a hyper-sphere by SVDD leads to avoid biasing toward major data. SVDD can gain the best description of data when its free parameters are set to proper values. In this paper, we propose belief-based chaotic krill herd algorithm for SVDD (BCKH-SVDD) with the aim of designing effective description of data. First, we introduce a new SVDD based on belief function theory, and then, we tune the free parameters by chaotic krill herd algorithm. Belief function theory is one of the best methods to enhance decision making for uncertain data. By adding a new belief-based weight, we can decide better about the data around the SVDD boundary and the classification will be more precise. Chaotic krill herd optimization algorithm introduces chaotic maps in the krill herd algorithm. With the help of chaotic maps, the two issues, namely local optima avoidance and convergence speed, can be overcome. Thus, chaotic krill herd algorithm is constructed based on chaotic functions and automatic switching between global and local searches of krill herd. To present the power of BCKH-SVDD, several experiments have been conducted based on tenfold cross-validation over real-world data sets from UCI repository. Experimental results show the superiority of the proposed algorithm to state-of-the-art methods in terms of classification accuracy, precision and recall measures.
KeywordsOne-class classification Support vector data description Belief function theory Outlier detection Belief-based chaotic algorithm for SVDD
Compliance with ethical standards
Conflict of interest
The authors declare that they have no conflict of interest.
This article does not contain any studies with animals performed by any of the authors.
- Cha M, Kim JS, Park SH, Baek J (2012) Nonparametric control chart using density weighted support vector data description. In: Proceedings of world academy of science, engineering and technology, world academy of science, engineering and technology (WASET) 1020Google Scholar
- El Boujnouni M, Jedra M, Zahid N (2014) Support vector domain description with a new confidence coefficient. In: 2014 9th International conference on intelligent systems: theories and applications (SITA-14), pp 1–8Google Scholar
- GhasemiGol M, Sabzekar M, Monsefi R, Naghibzadeh M, Yazdi HS (2010) A new support vector data description with fuzzy constraints. In: 2010 International conference on intelligent systems, modelling and simulation (ISMS), pp 10–14Google Scholar
- Hu Y, Liu JN, Wang Y, Lai L (2012) A weighted support vector data description based on rough neighborhood approximation. In: 2012 IEEE 12th international conference on data mining workshops (ICDMW), pp 635–642Google Scholar
- Kutsuna T, Yamamoto A (2016) Outlier detection using binary decision diagrams. Data Min Knowl Disc 458:1–26Google Scholar
- Lai V, Nguyen D, Nguyen K, Le T (2015) Mixture of support vector data description. In: 2nd National foundation for science and technology development conference on information and computer science (NICS), pp 135–140Google Scholar
- Lichman M (2013) UCI machine learning repository. http://archive.ics.uci.edu/ml. Accessed 10 Mar 2015
- Luo J, Ding L, Pan Z, Ni G, Hu G (2007) Research on cost-sensitive learning in one-class anomaly detection algorithms. In: Autonomic and trusted computing, pp 259–268Google Scholar
- Mirylenka K, Giannakopoulos G, Minh Do L, Palpanas T (2016) On classifier behavior in the presence of mislabeling noise. Data Min Knowl Disc. https://doi.org/10.1007/s10618-016-0484-8
- Peng X, Tan J (2015) EL-SVDD: an improved and localized multi-class classification algorithm. In: Applied mechanics and materials, Trans. Tech. Publ., vol 713, pp 1693–1698Google Scholar
- Sadeghi R, Hamidzadeh J (2016) Automatic support vector data description. Soft Comput. https://doi.org/10.1007/s00500-016-2317-5
- Tavakkoli A, Nicolescu M, Bebis G (2008) Incremental SVDD training: improving efficiency of background modeling in videos. In: Proceedings of the 10th IASTED international conference, pp 92Google Scholar
- Tax D, Laskov P (2003) Online SVM learning: from classification to data description and back. In: 2003 IEEE 13th workshop on neural networks for signal processing, 2003. NNSP’03, pp 499–508Google Scholar
- Theljani F, Laabidi K, Zidi S, Ksouri M (2015) Tennessee Eastman Process diagnosis based on dynamic classification with SVDD. J Dyn Syst Meas Control. https://doi.org/10.1115/1.4030429
- Wang L, Jia P, Huang T, Duan S, Yan J, Wang L (2016) A novel optimization technique to improve gas recognition by electronic noses based on the enhanced krill herd algorithm. Sensors 8:16Google Scholar