Abstract
Consider a finite collection \(\{\xi _k\}_{k=1}^p\) of vectors in the space \(\mathbb {R}^n\). The \(\xi _k\)’s are not to be seen as position points but as directions. This work addresses the problem of computing the ellipsoidal cone of minimal volume that contains all the \(\xi _k\)’s. The volume of an ellipsoidal cone is defined as the usual n-dimensional volume of a certain truncation of the cone. The central axis of the ellipsoidal cone of minimal volume serves to define the central direction of the datapoints, whereas its volume can be used as measure of dispersion of the datapoints.
Similar content being viewed by others
References
Ahipasaoglu, S.D., Sun, P., Todd, M.J.: Linear convergence of a modified Frank–Wolfe algorithm for computing minimum-volume enclosing ellipsoids. Optim. Methods Softw. 23, 5–19 (2008)
Astorino, A., Gaudioso, M., Seeger, A.: An illumination problem: optimal apex and optimal orientation for a cone of light. J. Glob. Optim. 58, 729–750 (2014)
Astorino, A., Gaudioso, M., Seeger, A.: Central axes and peripheral points in high dimensional directional datasets. Comput. Optim. Appl. (2015). doi:10.1007/s10589-014-9724-2
Ball, K.: An elementary introduction to modern convex geometry. In: Levy, S. (ed.) Flavors of Geometry, vol. 31, pp. 1–58. Math. Sci. Res. Inst. Publ., Cambridge Univ. Press, Cambridge (1997)
Berman, A.: Cones, matrices and mathematical programming. In: Lecture Notes in Economics and Mathematical Systems, vol. 79. Springer, Berlin (1973)
Danskin, J.M.: The theory of max–min, with applications. SIAM J. Appl. Math. 14, 641–664 (1966)
Goffin, J.-L.: The relaxation method for solving systems of linear inequalities. Math. Oper. Res. 5, 388–414 (1980)
Gourion, D., Seeger, A.: Solidity indices for convex cones. Positivity 16, 685–705 (2012)
Gruber, P.M.: John and Loewner ellipsoids. Discrete Comput. Geom. 46, 776–788 (2011)
Henrion, R., Seeger, A.: On properties of different notions of centers for convex cones. Set-Valued Var. Anal. 18, 205–231 (2010)
Henrion, R., Seeger, A.: Inradius and circumradius of various convex cones arising in applications. Set-Valued Var. Anal. 18, 483–511 (2010)
Iusem, A., Seeger, A.: Distances between closed convex cones: old and new results. J. Convex Anal. 17, 1033–1055 (2010)
Jerónimo-Castro, J., McAllister, T.B.: Two characterizations of ellipsoidal cones. J. Convex Anal. 20, 1181–1187 (2013)
Loewy, R., Schneider, H.: Positive operators on the n-dimensional ice cream cone. J. Math. Anal. Appl. 49, 375–392 (1975)
Rockafellar, R.T.: Convex Analysis. Princeton University Press, Princeton (1970)
Rockafellar, R.T., Wets, R .J.-B.: Variational Analysis. Springer, Berlin (1998)
Seeger, A., Torki, M.: Centers and partial volumes of convex cones. I. Basic theory. Beitr. Algebra Geom. 56, 227–248 (2015)
Seeger, A., Torki, M.: Conic version of Loewner–John ellipsoid theorem. Math. Program. 155, 403–433 (2016)
Stern, R., Wolkowicz, H.: Invariant ellipsoidal cones. Linear Algebra Appl. 150, 81–106 (1991)
Sun, P., Freund, R.M.: Computation of minimum-volume covering ellipsoids. Oper. Res. 52, 690–706 (2004)
Todd, M.J., Yildirim, E.A.: On Khachiyan’s algorithm for the computation of minimum-volume enclosing ellipsoids. Discrete Appl. Math. 155, 1731–1744 (2007)
Watson, G.S.: Statistics on Spheres. Wiley, New York (1983)
Acknowledgments
We would like to thank both referees for many helpful and constructive comments. The final form of this paper owes much to their recommendations.
Author information
Authors and Affiliations
Corresponding author
Additional information
José Vidal -Nuñez is supported by Consellería de la Educación de la Generalitat Valenciana, Spain, Pre-Doc Program Vali+d, DOCV 6791/07.06.2012, Grant ACIF-2013-156.
Appendix
Appendix
As mentioned already on a number of occasions, two different matrices \( B, C\in \mathbb {GL}(n)\) may produce the same ellipsoidal cone. The statement (d) in the next proposition provides an easily verifiable necessary condition for having the equality
A necessary and sufficient condition for (46) is given in Corollary 6.4. If P and D are convex cones in \(\mathbb {R}^n\), then \(Q \circeq P\) indicates that either \(Q=P\) or \(Q=-P\).
Proposition 6.3
For a pair of matrices \( B, C\in \mathbb {GL}(n)\), the following conditions are equivalent:
-
(a)
\(B^{-1}(\mathbb {L}_n) \circeq C^{-1}(\mathbb {L}_n)\).
-
(b)
\((BC^{-1})(\mathbb {L}_n) \circeq \mathbb {L}_n\).
-
(c)
\((BC^{-1})^\sharp \) is a positive multiple of \(J_n\).
-
(d)
\(B^\sharp \) is a positive multiple of \(C^\sharp \).
Proof
Clearly, (a) \(\Leftrightarrow \) (b) and (c) \(\Leftrightarrow \) (d). To prove the equivalence (b) \(\Leftrightarrow \) (c), there is no loss of generality in working with the particular case \(C=I_n\). So, we must check that
In fact, the above equivalence is a known result by Loewy and Schneider [14, Theorem 2.4]. We give below an alternative proof of (47) that is based on Proposition 3.3 and the spectral decomposition
Let \(B(\mathbb {L}_n) \circeq \mathbb {L}_n\). Hence, the ellipsoidal cone \(Q=B^{-1}(\mathbb {L}_n)\) is equal to \( \mathbb {L}_n\) or equal to \(-\mathbb {L}_n\). In particular, Q admits either \(u_n=e_n\) or \(u_n=-e_n\) as axial symmetry center. In addition, all the semiaxes lenghts of Q are equal to 1. From Proposition 3.3 we get \(\lambda _1= \ldots = \lambda _{n-1} =-\lambda _n\) and, a posteriori,
This shows that \(B^\sharp \) is a positive multiple of \(J_n\). Conversely, suppose that \(B^\sharp =\mu J_n \) for some \(\mu >0\). In such a case, \(u_n\in \{ e_n,-e_n\}\) and
Hence, the ellipsoidal cone \(Q=B^{-1}(\mathbb {L}_n)\) is symmetric with respect to the line generated by \(e_n\) and, thanks to (48) and Proposition 3.3, all the semi-axes lengths of Q are equal to 1. In conclusion, Q is either the Lorentz cone or the opposite of the Lorentz cone. \(\square \)
Corollary 6.4
Consider a pair \( B, C\in \mathbb {M}_n\) of invertible matrices and write
with \(t\in \mathbb {R}, M\in \mathbb {M}_{n-1}\), and \(z,q\in \mathbb {R}^{n-1}\). Then
Proof
By using Loewy and Schneider [14, Theorem 2.4] or the equivalence (b) \(\Leftrightarrow \) (c) in Proposition 6.3, we get
The case \(BC^{-1}(\mathbb {L}_n)= \mathbb {L}_n\) occurs when the last entry of \( BC^{-1}e_n \) is positive, i.e., when \(t>0\). \(\square \)
Rights and permissions
About this article
Cite this article
Seeger, A., Vidal -Nuñez, J. Measuring centrality and dispersion in directional datasets: the ellipsoidal cone covering approach. J Glob Optim 68, 279–306 (2017). https://doi.org/10.1007/s10898-016-0464-y
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10898-016-0464-y