Minkowski-type theorems and least-squares clustering

F. Aurenhammer, F. Hoffmann, Boris Aronov

    Research output: Contribution to journalArticle

    Abstract

    Dissecting Euclidean d-space with the power diagram of n weighted point sites partitions a given m-point set into clusters, one cluster for each region of the diagram. In this manner, an assignment of points to sites is induced. We show the equivalence of such assignments to constrained Euclidean least-squares assignments. As a corollary, there always exists a power diagram whose regions partition a given d-dimensional m-point set into clusters of prescribed sizes, no matter where the sites are placed. Another consequence is that constrained least-squares assignments can be computed by finding suitable weights for the sites. In the plane, this takes roughly O(n2m) time and optimal space O(m), which improves on previous methods. We further show that a constrained least-squares assignment can be computed by solving a specially structured linear program in n + 1 dimensions. This leads to an algorithm for iteratively improving the weights, based on the gradient-descent method. Besides having the obvious optimization property, least-squares assignments are shown to be useful in solving a certain transportation problem, and in finding a least-squares fitting of two point sets where translation and scaling are allowed. Finally, we extend the concept of a constrained least-squares assignment to continuous distributions of points, thereby obtaining existence results for power diagrams with prescribed region volumes. These results are related to Minkowski's theorem for convex polytopes. The aforementioned iterative method for approximating the desired power diagram applies to continuous distributions as well.

    Original languageEnglish (US)
    Pages (from-to)61-76
    Number of pages16
    JournalAlgorithmica (New York)
    Volume20
    Issue number1
    StatePublished - 1998

    Fingerprint

    Iterative methods
    Power Diagram
    Least Squares
    Assignment
    Clustering
    Constrained Least Squares
    Theorem
    Point Sets
    Continuous Distributions
    Euclidean
    Partition
    Gradient Descent Method
    Convex Polytopes
    Least Square Fitting
    Transportation Problem
    D-space
    Linear Program
    Existence Results
    Corollary
    Diagram

    Keywords

    • Least-squares clustering
    • Point partitioning
    • Power diagrams

    ASJC Scopus subject areas

    • Computer Graphics and Computer-Aided Design
    • Software
    • Safety, Risk, Reliability and Quality
    • Applied Mathematics

    Cite this

    Aurenhammer, F., Hoffmann, F., & Aronov, B. (1998). Minkowski-type theorems and least-squares clustering. Algorithmica (New York), 20(1), 61-76.

    Minkowski-type theorems and least-squares clustering. / Aurenhammer, F.; Hoffmann, F.; Aronov, Boris.

    In: Algorithmica (New York), Vol. 20, No. 1, 1998, p. 61-76.

    Research output: Contribution to journalArticle

    Aurenhammer, F, Hoffmann, F & Aronov, B 1998, 'Minkowski-type theorems and least-squares clustering', Algorithmica (New York), vol. 20, no. 1, pp. 61-76.
    Aurenhammer F, Hoffmann F, Aronov B. Minkowski-type theorems and least-squares clustering. Algorithmica (New York). 1998;20(1):61-76.
    Aurenhammer, F. ; Hoffmann, F. ; Aronov, Boris. / Minkowski-type theorems and least-squares clustering. In: Algorithmica (New York). 1998 ; Vol. 20, No. 1. pp. 61-76.
    @article{ee3806df69c34698908eb7e3ae4bfae9,
    title = "Minkowski-type theorems and least-squares clustering",
    abstract = "Dissecting Euclidean d-space with the power diagram of n weighted point sites partitions a given m-point set into clusters, one cluster for each region of the diagram. In this manner, an assignment of points to sites is induced. We show the equivalence of such assignments to constrained Euclidean least-squares assignments. As a corollary, there always exists a power diagram whose regions partition a given d-dimensional m-point set into clusters of prescribed sizes, no matter where the sites are placed. Another consequence is that constrained least-squares assignments can be computed by finding suitable weights for the sites. In the plane, this takes roughly O(n2m) time and optimal space O(m), which improves on previous methods. We further show that a constrained least-squares assignment can be computed by solving a specially structured linear program in n + 1 dimensions. This leads to an algorithm for iteratively improving the weights, based on the gradient-descent method. Besides having the obvious optimization property, least-squares assignments are shown to be useful in solving a certain transportation problem, and in finding a least-squares fitting of two point sets where translation and scaling are allowed. Finally, we extend the concept of a constrained least-squares assignment to continuous distributions of points, thereby obtaining existence results for power diagrams with prescribed region volumes. These results are related to Minkowski's theorem for convex polytopes. The aforementioned iterative method for approximating the desired power diagram applies to continuous distributions as well.",
    keywords = "Least-squares clustering, Point partitioning, Power diagrams",
    author = "F. Aurenhammer and F. Hoffmann and Boris Aronov",
    year = "1998",
    language = "English (US)",
    volume = "20",
    pages = "61--76",
    journal = "Algorithmica",
    issn = "0178-4617",
    publisher = "Springer New York",
    number = "1",

    }

    TY - JOUR

    T1 - Minkowski-type theorems and least-squares clustering

    AU - Aurenhammer, F.

    AU - Hoffmann, F.

    AU - Aronov, Boris

    PY - 1998

    Y1 - 1998

    N2 - Dissecting Euclidean d-space with the power diagram of n weighted point sites partitions a given m-point set into clusters, one cluster for each region of the diagram. In this manner, an assignment of points to sites is induced. We show the equivalence of such assignments to constrained Euclidean least-squares assignments. As a corollary, there always exists a power diagram whose regions partition a given d-dimensional m-point set into clusters of prescribed sizes, no matter where the sites are placed. Another consequence is that constrained least-squares assignments can be computed by finding suitable weights for the sites. In the plane, this takes roughly O(n2m) time and optimal space O(m), which improves on previous methods. We further show that a constrained least-squares assignment can be computed by solving a specially structured linear program in n + 1 dimensions. This leads to an algorithm for iteratively improving the weights, based on the gradient-descent method. Besides having the obvious optimization property, least-squares assignments are shown to be useful in solving a certain transportation problem, and in finding a least-squares fitting of two point sets where translation and scaling are allowed. Finally, we extend the concept of a constrained least-squares assignment to continuous distributions of points, thereby obtaining existence results for power diagrams with prescribed region volumes. These results are related to Minkowski's theorem for convex polytopes. The aforementioned iterative method for approximating the desired power diagram applies to continuous distributions as well.

    AB - Dissecting Euclidean d-space with the power diagram of n weighted point sites partitions a given m-point set into clusters, one cluster for each region of the diagram. In this manner, an assignment of points to sites is induced. We show the equivalence of such assignments to constrained Euclidean least-squares assignments. As a corollary, there always exists a power diagram whose regions partition a given d-dimensional m-point set into clusters of prescribed sizes, no matter where the sites are placed. Another consequence is that constrained least-squares assignments can be computed by finding suitable weights for the sites. In the plane, this takes roughly O(n2m) time and optimal space O(m), which improves on previous methods. We further show that a constrained least-squares assignment can be computed by solving a specially structured linear program in n + 1 dimensions. This leads to an algorithm for iteratively improving the weights, based on the gradient-descent method. Besides having the obvious optimization property, least-squares assignments are shown to be useful in solving a certain transportation problem, and in finding a least-squares fitting of two point sets where translation and scaling are allowed. Finally, we extend the concept of a constrained least-squares assignment to continuous distributions of points, thereby obtaining existence results for power diagrams with prescribed region volumes. These results are related to Minkowski's theorem for convex polytopes. The aforementioned iterative method for approximating the desired power diagram applies to continuous distributions as well.

    KW - Least-squares clustering

    KW - Point partitioning

    KW - Power diagrams

    UR - http://www.scopus.com/inward/record.url?scp=3843098445&partnerID=8YFLogxK

    UR - http://www.scopus.com/inward/citedby.url?scp=3843098445&partnerID=8YFLogxK

    M3 - Article

    AN - SCOPUS:3843098445

    VL - 20

    SP - 61

    EP - 76

    JO - Algorithmica

    JF - Algorithmica

    SN - 0178-4617

    IS - 1

    ER -