Interval Chi-Square Score (ICSS): Feature Selection of Interval Valued Data
In this paper, a novel feature ranking criterion suitable for interval valued feature selection is proposed. The proposed criterion simulates the characteristics of the well known statistical criterion - chi-square in selecting the interval valued features effectively and hence called as Interval Chi-Square Score. Moreover, the paper also highlights the alternative approach proposed for computing the frequency of the distribution of interval valued data. For experimentation purpose, two standard benchmarking interval valued datasets are used with a suitable symbolic classifier for classification. The performance of the proposed ranking criterion is evaluated in terms of accuracy and the results are comparatively better than the contemporary interval valued feature selection methods.
KeywordsInterval valued data Symbolic feature selection Chi-square score
The author N Vinay Kumar acknowledges the Department of Science & Technology, Govt. of India for their financial support rendered in terms of DST-INSPIRE fellowship.
- 3.Duda, O.R., Hart, E.P., Stork, G.D.: Pattern Classification, 2nd edn. Wiley-Interscience (2000)Google Scholar
- 5.Guru, D.S., Nagendraswamy, H.S.: Symbolic representation and classification of two- dimensional shapes. In: Proceedings of the 3rd Workshop on Computer Vision, Graphics, and Image Processing (WCVGIP), pp. 19–24 (2006)Google Scholar
- 6.Guru, D.S. Vinay Kumar, N.: Class specific feature selection for interval valued data through interval K-means clustering, RTIP2R 2016, CCIS, vol. 709, pp. 228–239. Springer (2017)Google Scholar
- 9.Guru, D.S., Vinay Kumar, N.: Novel feature ranking criteria for interval valued feature selection. In: Proceedings of the IEEE International Conference on Advances in Computing, Communications and Informatics, pp. 149–155 (2016)Google Scholar
- 12.Hsiao, C.C., Chuang, C.C., Su, S.F.: Robust Gaussian Kernel based approach for feature selection. In: Advances in Intelligent Systems and Computing, vol. 268, pp. 25–33 (2014)Google Scholar
- 13.Ichino, M.: Feature selection for symbolic data classification. In: New Approaches in Classification and Data Analysis, pp. 423–429. Springer (1994). section 2Google Scholar
- 14.Kiranagi, B.B., Guru D.S., Ichino, M.: Exploitation of multivalued type proximity for symbolic feature selection. In: Proceeding of the Internal Conference on Computing: Theory and Applications, pp. 320–324. IEEE (2007)Google Scholar
- 15.Liu, Q., Wang, J., Xiao, J., Zhu, H.: Mutual information based feature selection for symbolic interval data. In: Proceedings of International Conference on Software Intelligence, Technologies and Applications, pp. 62–69 (2014)Google Scholar
- 16.Liu, H., Setiono, R.: Chi2: feature selection and discretization of numeric attributes. In: Proceedings of 7th IEEE International Conference on Tools with Artificial Intelligence, Technologies and Applications, pp. 388–391 (1995)Google Scholar
- 18.Vinay Kumar, N., Guru, D.S.: A novel feature ranking criterion for supervised interval valued feature selection for classification. In: Proceedings of the 14th IAPR International Conference on Document Analysis and Recognition (ICDAR), pp. 71–76 (2017)Google Scholar