Effective strategies for segmenting data into coherent subsetsReport as inadecuate

Effective strategies for segmenting data into coherent subsets - Download this document for free, or read online. Document in PDF available to download.

1 SELECT - Model selection in statistical learning Inria Saclay - Ile de France, LMO - Laboratoire de Mathématiques d-Orsay, CNRS - Centre National de la Recherche Scientifique : UMR

Abstract : Automatic segmentation of data into coherent subsets is important in applications as varied as signal processing, bioinformatics and pharmacology. Under this general framework, we investigate the problem of data-driven reconstruction of an unknown, piecewise-constant density function and propose two methods to solve it; the first is directly inspired by the segmentation approach, whereas the second uses a maximum likelihood approach. Motivated by a problem in pharmacometrics, we then introduce a segmentation algorithm which fits into the same general framework and is used for automatically binning data for model assessment purposes.

Keywords : Histogram Segmentation Signal Piecewise-constant density function Maximum likelihood Pharmacology Visual Predictive Check Dynamic programming Histogram.

Author: Kevin Bleakley - Marc Lavielle -

Source: https://hal.archives-ouvertes.fr/


Related documents