A0786
Title: Asymptotic behaviors of k-means under high dimensional settings
Authors: Kento Egashira - Tokyo University of Science (Japan) [presenting]
Kazuyoshi Yata - University of Tsukuba (Japan)
Makoto Aoshima - University of Tsukuba (Japan)
Abstract: While k-means has been approved as a useful methodology for analysing gene expression microarray data on behalf of high-dimensional, low-sample-size (HDLSS), k-means is not sufficiently studied theoretically under high-dimensional settings. The asymptotic properties of k-means are proved under mild and practical settings for HDLSS data. Results for k-means were also applied to kernel k-means. The current comprehension of k-means proceeds. Finally, numerical simulation studies are given, and the performance of the k-means for high-dimensional data is discussed.
