Fidelity-commensurability tradeoff in joint embedding of disparate dissimilarities

DOI10.1007/S00357-016-9214-6MaRDI QIDQ2628070zbMATH OpenOpenAlexFDO

Authors Sancar Adali, Carey E. Priebe

Publication date 12 June 2017

Published in Journal of Classification (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1306.1977

zbMATH Keywords

embedding of dissimilaritiesinferencejoint embeddingmultidimensional scalingmultimodal data fusion

Mathematics Subject Classification ID

Classification and discrimination; cluster analysis (statistical aspects) (62H30)

Abstract: In various data settings, it is necessary to compare observations from disparate data sources. We assume the data is in the dissimilarity representation and investigate a joint embedding method that results in a commensurate representation of disparate dissimilarities. We further assume that there are "matched" observations from different conditions which can be considered to be highly similar, for the sake of inference. The joint embedding results in the joint optimization of fidelity (preservation of within-condition dissimilarities) and commensurability (preservation of between-condition dissimilarities between matched observations). We show that the tradeoff between these two criteria can be made explicit using weighted raw stress as the objective function for multidimensional scaling. In our investigations, we use a weight parameter, $w$, to control the tradeoff, and choose match detection as the inference task. Our results show weights that are optimal (with respect to the inference task) are different than equal weights for commensurability and fidelity and the proposed weighted embedding scheme provides significant improvements in statistical power.

Recommendations

• Manifold matching: joint optimization of fidelity and commensurability
• Dimensionality reduction on the Cartesian product of embeddings of multiple dissimilarity matrices
• Structure preserving embedding of dissimilarity data
• Encoding dissimilarity data for statistical model building
• scientific article; zbMATH DE number 756189

Cites work

• Analysis of individual differences in multidimensional scaling via an \(n\)-way generalization of ``Eckart-Young decomposition
• Asymptotics for \(M\)-estimators defined by convex minimization
• Automatic dimensionality selection from the scree plot via the use of profile likelihood
• Canonical Correlation Analysis: An Overview with Application to Learning Methods
• scientific article; zbMATH DE number 3605829 (Why is no real title available?)
• scientific article; zbMATH DE number 3385438 (Why is no real title available?)
• Manifold matching: joint optimization of fidelity and commensurability
• Modern multidimensional scaling: theory and applications
• Multidimensional scaling. I: Theory and method
• RELATIONS BETWEEN TWO SETS OF VARIATES
• The Dissimilarity Representation for Pattern Recognition

Cited in

(2)

• Dimensionality reduction on the Cartesian product of embeddings of multiple dissimilarity matrices
• Manifold matching: joint optimization of fidelity and commensurability

This page was built for publication: Fidelity-commensurability tradeoff in joint embedding of disparate dissimilarities

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2628070)