Abstract:
The problem of multidimensional object classification with small training sample is considered. The following algorithms of estimating variable informativeness are considered: Ad, Del, AdDel.
A new algorithm for selecting informative variables is proposed. It is based on the optimization of the coefficient vector of the kernel fuzziness. Some modification of this algorithm is also discussed.
The comparative analysis of existing methods for selecting informative variables is presented.
Keywords:
classification, small training sample, informative variable, optimization of the coefficient vector of the kernel fuzziness.
The study was performed by a grant from the Russian Science Foundation (project no. 16-19-10089).
Received: 23.06.2016 Received in revised form: 14.08.2016 Accepted: 14.09.2016
Bibliographic databases:
Document Type:
Article
UDC:519.87
Language: English
Citation:
Eugene D. Mihov, Oleg V. Nepomnyashchiy, “Selecting informative variables in the identification problem”, J. Sib. Fed. Univ. Math. Phys., 9:4 (2016), 473–480