New Depth-Based Modification of the k-Nearest Neighbour Method

Result code in IS VaVaI
<a href="https://www.isvavai.cz/riv?ss=detail&h=RIV%2F61989592%3A15310%2F14%3A33158321" target="_blank" >RIV/61989592:15310/14:33158321 - isvavai.cz</a>
Result on the web
<a href="http://www.scionlinepublishing.org/journals/STSA/papers/download/2105-632.pdf" target="_blank" >http://www.scionlinepublishing.org/journals/STSA/papers/download/2105-632.pdf</a>
DOI - Digital Object Identifier
—

Result language
angličtina
Original language name
New Depth-Based Modification of the k-Nearest Neighbour Method
Original language description
In this paper, we propse a new depth-based modification of the k-nearest neighbor method. We want to classify a new d-dimensional observation into one of two groups. Newly proposed method is based on the idea that points with similar location with respect to the two considered distributions will have similar depths with respect to the distributions. When classifying a new object, we first calculate its depth to each group of points in the training set (we reduce the dimension: R^d to R^2) and then we can find k-nearest neighbors (points of the training set with similar pair of depths to the considered distributions). Good properties of newly proposed classifier are shown in a short simulation study.
Czech name
—
Czech description
—

Type
J<sub>x</sub> - Unclassified - Peer-reviewed scientific article (Jimp, Jsc and Jost)
CEP classification
BA - General mathematics
OECD FORD branch
—

Project
<a href="/en/project/EE2.3.20.0170" target="_blank" >EE2.3.20.0170: Building of Research Team in the Field of Environmental Modeling and the Use of Geoinformation Systems with the Consequence in Participation in International Networks and Programs</a><br>
Continuities
P - Projekt vyzkumu a vyvoje financovany z verejnych zdroju (s odkazem do CEP)

Publication year
2014
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů

Similar results(10)