scispace - formally typeset
Search or ask a question
Topic

Mahalanobis distance

About: Mahalanobis distance is a research topic. Over the lifetime, 4616 publications have been published within this topic receiving 95294 citations.


Papers
More filters
Journal ArticleDOI
TL;DR: A simple guide to the calculation of Mahalanobis' D2 statistic, intended for those with limited mathematical background and no access to computing equipment, is presented in hopes that such a procedure will permit application of this statistic, by anthropologists, with a minimum of effort.
Abstract: It is suggested that, while use of many statistical techniques is facilitated by electronic aids, a computer is not always a necessity. A simple guide to the calculation of Mahalanobis' D2 statistic, intended for those with limited mathematical background and no access to computing equipment, is presented in hopes that such a procedure will permit application of this statistic, by anthropologists, with a minimum of effort.

25 citations

Journal Article
TL;DR: In this paper, Wang et al. improved traditional grey target decision methods with the approach of weighted Mahalanobis distance and discussed the property of the improved grey target method, taking into account the correlation among different indexes as well as the influences of both different dimensions and differences of importance on effect of decisions.
Abstract: Taking into account the correlation among different indexes as well as the influences of both different dimensions and differences of importance on effect of decisions,this paper deals with improving traditional grey target decision methods with the approach of weighted Mahalanobis distance and discusses the property of the improved grey target method.The results indicate that the improved grey target could meet the invariance of effect samples after nonsingular linear transformation caused by clout distance;the weighted Mahalanobis distance is just the weighted Euclidean distance when there is no correlation among observed indexes,which avoids correlation of decision indexes,influences of different dimension and differences of importance on effect of decisions,and the incompatibility of grey target transformation.Finally,the effectiveness and availability of the model are proved by a real case in point.

25 citations

Journal ArticleDOI
TL;DR: A novel robust Bayesian network is proposed for process modeling with low-quality data since unreliable data can cause model parameters to deviate from the real distributions and make network structures unable to characterize the true causalities.

25 citations

Journal ArticleDOI
TL;DR: A two-stage MCS classification approach, coupled with Binary Particle Swarm Optimization, is proposed to optimize the process of selecting the most significant features and to search for the optimal decision boundary to discriminate healthy and unhealthy components.

25 citations

Journal ArticleDOI
TL;DR: This article first construct Mahalanobis distance in the kernel space and then proposes a novel fuzzy clustering model with a kernelized MahalanOBis distance, namely KMD-FC, which outperformed the state-of-the-art methods in comparison.
Abstract: Data samples of complicated geometry and nonlinear separability are considered as common challenges to clustering algorithms. In this article, we first construct Mahalanobis distance in the kernel space and then propose a novel fuzzy clustering model with a kernelized Mahalanobis distance, namely KMD-FC. The key contributions of KMD-FC include: first, the construction of KMD matrix is innovatively transformed from the Euclidean distance kernel matrix, which is able to effectively avoid the problem of “curse of dimensionality” posed by explicitly calculating the sample covariance matrix in the kernel space; second, for the first time, the kernelized Gustafson–Kessel (GK) fuzzy C-means algorithm is achieved, which is critically important to extend the applications of the GK algorithm to the nonlinear classification tasks; finally, taking account of the overall distribution of samples in the kernel space after kernel mapping to improve the generalizability of the proposed KMD-FC clustering method. Comprehensive experiments conducted on a wide range of datasets, including synthetic datasets and machine learning repository (UCI) datasets, have validated that the proposed clustering algorithm outperformed the state-of-the-art methods in comparison.

25 citations


Network Information
Related Topics (5)
Cluster analysis
146.5K papers, 2.9M citations
79% related
Artificial neural network
207K papers, 4.5M citations
79% related
Feature extraction
111.8K papers, 2.1M citations
77% related
Convolutional neural network
74.7K papers, 2M citations
77% related
Image processing
229.9K papers, 3.5M citations
76% related
Performance
Metrics
No. of papers in the topic in previous years
YearPapers
20241
2023208
2022452
2021232
2020239
2019249