Robust nonparametric probability density estimation by soft clustering

 

Abstract

A method to estimate the probability density function of multivariate distributions is presented. The classical Parzen window approach builds a spherical Gaussian density around every input sample. This choice of the kernel density yields poor robustness for real input datasets. We use multivariate Student-t distributions in order to improve the adaptation capability of the model. Our method has a first stage where hard neighbourhoods are determined for every sample. Then soft clusters are considered to merge the information coming from several hard neighbourhoods. Hence, a specific mixture component is learned for each soft cluster. This leads to outperform other proposals where the local kernel is not as robust and/or there are no smoothing strategies, like the manifold Parzen windows. © Springer-Verlag Berlin Heidelberg 2008.

Cites

The following graph plots the number of cites received by this work from its publication, on a yearly basis.

Citation

Please, cite this work as:

[Lóp+08] E. López-Rubio, J. Ortiz-De-Lazcano-Lobato, D. López-Rodríguez, et al. “Robust nonparametric probability density estimation by soft clustering”. In: Artificial Neural Networks - ICANN 2008 , 18th International Conference, Prague, Czech Republic, September 3-6, 2008, Proceedings, Part I. Ed. by V. K. á, R. Neruda and J. Koutník. Vol. 5163 LNCS. Lecture Notes in Computer Science PART 1. cited By 0; Conference of 18th International Conference on Artificial Neural Networks, ICANN 2008 ; Conference Date: 3 September 2008 Through 6 September 2008; Conference Code:73798. Prague: Springer, 2008, pp. 155-164. DOI: 10.1007/978-3-540-87536-9_17. URL: https://doi.org/10.1007/978-3-540-87536-9_17.

BibTeX

@InProceedings{LopezRubio2008,
     author = {E. López-Rubio and J.M. Ortiz-De-Lazcano-Lobato and D. López-Rodríguez and M. {Del Carmen Vargas-Gonzalez}},
     booktitle = {Artificial Neural Networks - {ICANN} 2008 , 18th International Conference, Prague, Czech Republic, September 3-6, 2008, Proceedings, Part {I}},
     title = {Robust nonparametric probability density estimation by soft clustering},
     year = {2008},
     address = {Prague},
     editor = {Vera Kurko á and Roman Neruda and Jan Koutník},
     note = {cited By 0; Conference of 18th International Conference on Artificial Neural Networks, ICANN 2008 ; Conference Date: 3 September 2008 Through 6 September 2008; Conference Code:73798},
     number = {PART 1},
     pages = {155-164},
     publisher = {Springer},
     series = {Lecture Notes in Computer Science},
     volume = {5163 LNCS},
     abstract = {A method to estimate the probability density function of multivariate distributions is presented. The classical Parzen window approach builds a spherical Gaussian density around every input sample. This choice of the kernel density yields poor robustness for real input datasets. We use multivariate Student-t distributions in order to improve the adaptation capability of the model. Our method has a first stage where hard neighbourhoods are determined for every sample. Then soft clusters are considered to merge the information coming from several hard neighbourhoods. Hence, a specific mixture component is learned for each soft cluster. This leads to outperform other proposals where the local kernel is not as robust and/or there are no smoothing strategies, like the manifold Parzen windows. © Springer-Verlag Berlin Heidelberg 2008.},
     bibsource = {dblp computer science bibliography, https://dblp.org},
     biburl = {https://dblp.org/rec/conf/icann/Lopez-RubioOLV08.bib},
     document_type = {Conference Paper},
     doi = {10.1007/978-3-540-87536-9_17},
     journal = {Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)},
     keywords = {Data sets; Gaussian density; Input sample; Kernel density; Local kernel; Mixture components; Multivariate distributions; Multivariate Student; Non-parametric; Parzen windows; Probability density estimation; Soft clustering, Backpropagation; Neural networks; Probability distributions; Windows, Probability density function},
     source = {Scopus},
     url = {https://doi.org/10.1007/978-3-540-87536-9_17},
}