Abstract
Mixture models constitute a popular type of probabilistic neural networks which model the density of a dataset using a convex combination of statistical distributions, with the Gaussian distribution being the one most commonly used. In this work we propose a new probability density function, called the Π-sigmoid, from its ability to form the shape of the letter “Π” by appropriately combining two sigmoid functions. We demonstrate its modeling properties and the different shapes that can take for particular values of its parameters. We then present the Π-sigmoid mixture model and propose a maximum likelihood estimation method to estimate the parameters of such a mixture model using the Generalized Expectation Maximization algorithm. We assess the performance of the proposed method using synthetic datasets and also on image segmentation and illustrate its superiority over Gaussian mixture models.
Chapter PDF
Similar content being viewed by others
Keywords
References
Pelleg, D., Moore, A.: Mixture of rectangles: Interpretable soft clustering. In: Proc. ICML (2001)
Vlassis, N., Likas, A.: A greedy EM algorithm for Gaussian mixture learning. Neural Processing Letters 15, 77–87 (2002)
McLachlan, J.G., Krishnan, T.: Finite Mixture Models. Wiley, Chichester (2000)
McLachlan, J.G., Krishnan, T.: The EM algorithm and extensions. Marcel Dekker, New York (1997)
Author information
Authors and Affiliations
Editor information
Rights and permissions
Copyright information
© 2008 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Alivanoglou, A., Likas, A. (2008). Probabilistic Models Based on the Π-Sigmoid Distribution. In: Prevost, L., Marinai, S., Schwenker, F. (eds) Artificial Neural Networks in Pattern Recognition. ANNPR 2008. Lecture Notes in Computer Science(), vol 5064. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-540-69939-2_4
Download citation
DOI: https://doi.org/10.1007/978-3-540-69939-2_4
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-540-69938-5
Online ISBN: 978-3-540-69939-2
eBook Packages: Computer ScienceComputer Science (R0)