6533b873fe1ef96bd12d558a

RESEARCH PRODUCT

Measuring Dissimilarity Between Curves by Means of Their Granulometric Size Distributions

Teresa LeónFermin MallorMartín GastónGuillermo Ayala

subject

Functional principal component analysisPointwiseDynamic time warpingComputer sciencebusiness.industryFunctional data analysisPattern recognitionMathematical morphologyMeasure (mathematics)ToolboxComputingMethodologies_PATTERNRECOGNITIONArtificial intelligenceCluster analysisbusiness

description

The choice of a dissimilarity measure between curves is a key point for clustering functional data. Functions are usually pointwise compared and, in many situations, this approach is not appropriate. Mathematical Morphology provides us with a toolbox to overcome this problem. We propose some dissimilarity measures based on morphological granulometries and their performance is evaluated on some functional datasets.

yearjournalcountryeditionlanguage
2008-01-01
https://doi.org/10.1007/978-3-7908-2062-1_6