Cybernetics and Systems Analysis / Issue (2016, 52 (2))
Galkin O.A.
Affine-invariant classifier of extrapolation depth based on multilevel smoothing structure A nonparametric affine-invariant extrapolation depth-based classifier resistant to spikes and extreme values is proposed and investigated. A multilevel smoothing structure is proposed that makes it possible to obtain global properties of density functions and class boundaries under appropriate regularity conditions. The extrapolation depth-based classifier uses kernel density estimates to efficiently classify multidimensional data at different smoothing levels. © 2016, Springer Science+Business Media New York. Keywords: depth function, kernel density estimate, smoothing level, Computer science, Cybernetics, Affine invariant, Class boundary, Global properties, Kernel density estimate, Multidimensional data, Non-parametric, Regularity condition, smoothing level, Extrapolation
Cite: Galkin O.A.
(2016). Affine-invariant classifier of extrapolation depth based on multilevel smoothing structure. Cybernetics and Systems Analysis, 52 (2), 64-72. doi: https://doi.org/10.1007/s10559-016-9819-0 http://jnas.nbuv.gov.ua/article/UJRN-0000496948 [In Russian]. |