Distance-based classifier by data transformation for high-dimension, strongly spiked eigenvalue models (Q2000734)

This paper proposes distance-based classifiers by the eigenstructures estimation in high-dimensional settings. The efficiency of the proposed classifier is demonstrated using medical datasets. It is to be noted that the random forests behave well in high-dimensional settings. Readers may refer to [\textit{A. V. Zhukov} and \textit{D. N. Sidorov}, Vestn. Yuzhno-Ural. Gos. Univ., Ser. Mat. Model. Program. 9, No. 4, 86--95 (2016; Zbl 1442.68210)] for a modification to handle the concept drifting.