Abstract
When analyzing high-dimensional data, it is often appropriate to pay attention only to the direction of each datum, disregarding its norm. The von Mises–Fisher (vMF) distribution is a natural probability distribution for such data. When we estimate the parameters of vMF distributions, parameter κ which corresponds to the degree of concentration is difficult to obtain, and some approximations are necessary. In this article, we propose an iterative algorithm using fixed points to obtain the maximum likelihood estimate (m.l.e.) for κ. We prove that there is a unique local maximum for κ. Besides, using a specific function to calculate the m.l.e., we obtain the upper and lower bounds of the interval in which the exact m.l.e. exists. In addition, based on these bounds, a new and good approximation is derived. The results of numerical experiments demonstrate the new approximation exhibits higher precision than traditional ones.
Similar content being viewed by others
References
Abramowitz M, Stegun IA (1972) Handbook of mathematical functions, whith formulas, graphs, and mathematical tables. Courier Dover Publications
Agarwal RP, Meehan M, O’Regan D (2001) Fixed Point Theory and Applications. Cambridge University Press, Cambridge
Aitken AC (1926) On bernoulli’s numerical solution of algebraic equations. Proc R Soc Edinb 46:289–305
Amos DE (1974) Computation of modified bessel function and their ratios. Math Comput 28: 239–251
Banerjee A, Dhillon IS, Gosh J, Sra S (eds) (2003) Generative model-based clustering of directional data. In: proceedings of the ninth ACM SIGKEE international conference on knowledge discovery and data mining
Banerjee A, Dhillon IS, Ghosh J, Sra S (2005) Clustering on the unit hypersphere using von mises-fisher distribution. J Mach Learn Res 6:1345–1382
Devroye L (2002) Simulating bessel random variables. Stat Probab Lett 57(3):249–257
Dhillon IS, Sra S (2003) Modeling data using directional distributions, Technical report. UTCS technical report
Eisen MB, Spellman PT, Brown PO, Botstein D (1998) Cluster analysis and display of genome-wide expression patterns. Proc Natl Acad Sci USA 95:14863–14868
Fisher NI, Lewis T, Embleton BJJ (1987) Statistical analysis of spherical data. Cambridge University Press, Cambridge
Hill GW (1981) Evaluation and inversion of the ratios of modified bessel functions, \({i_1(x)/i_0(x)}\) and \({i_{1.5}(x)/i_{0.5}(x)}\) . ACM Trans Math Softw 7(2):199–208
Jost J (1997) Postmodern analysis. Springer, Heidelberg
Mardia KV, Jupp P (2000) Directional statistics, 2nd (edn.) Wiley, New York
Nasell I (1974) Inequalities for modified bessel functions. Math Comput 28(125):253–256
Thiruvenkatachar VK, Najundiah TS (1951) Inequalities concerning bessel functions and orthogonal polynomials. Proc Indian Nat Sci Acad Part A 33:373–384
Watson GN (1995) A treatise on the theory of bessel Functions. (2nd edn.) Cambridge University Press, Cambridge
Yuan L, Kalbfleisch JD (2000) On the bessel distribution and related problems. Ann Insit Stat Math 52(3):438–447
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Tanabe, A., Fukumizu, K., Oba, S. et al. Parameter estimation for von Mises–Fisher distributions. Computational Statistics 22, 145–157 (2007). https://doi.org/10.1007/s00180-007-0030-7
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00180-007-0030-7