A novel neighbor selection approach for KNN: a physiological status prediction case study

Authors:
Fatemeh Kaveh-Yazdy;Mohammad-Reza Zare-Mirakabad;Feng Xia
Affiliations:
Yazd University, Yazd, Iran;Yazd University, Yazd, Iran;Dalian University of Technology, Dalian, China
Venue:
Proceedings of the 1st International Workshop on Context Discovery and Data Mining
Year:
2012

Citing 7
Cited 0

Interestingness measures for data mining: A survey

ACM Computing Surveys (CSUR)
Wearable wireless sensor network to assess clinical status in patients with neurological disorders

Proceedings of the 6th international conference on Information processing in sensor networks
Pervasive healthcare and wireless health monitoring

Mobile Networks and Applications
SRDA: An Efficient Algorithm for Large-Scale Discriminant Analysis

IEEE Transactions on Knowledge and Data Engineering
KNN-kernel density-based clustering for high-dimensional multivariate data

Computational Statistics & Data Analysis
Data Mining with Decision Trees: Theroy and Applications

Data Mining with Decision Trees: Theroy and Applications
Triggers and Monitoring in Intelligent Personal Health Record

Journal of Medical Systems

Quantified Score

Hi-index	0.00

Visualization

Abstract

Conventional weighting KNNs enhance the accuracy of label selection by weighting the neighbors. In this paper, however, we propose a novel weighting approach which weights the distances to find the neighbors more accurately. We take importance of size of classes and dispersity of samples into account for this purpose. Moreover we use LDA that saves discrimination level of data for reducing dimension of problem space. We show the effectivity of our proposed method on PDMC-04 dataset to predict physiological status of human subjects.