Title :
Mode detection by relaxation
Author :
Touzani, A. ; Postaire, J.-G.
Author_Institution :
Centre d´´Autom., Univ. des Sci. et Tech. de Lille Flandres-Artois, Villeneuve, d´´Ascq, France
fDate :
11/1/1988 12:00:00 AM
Abstract :
A relaxation process is described and applied to the detection of the modes of the underlying probability density function for clustering multidimensional data. When the process is iterated, the label probabilities become very high in the modal region and very low in the valleys, so that mode detection becomes trivial
Keywords :
computerised pattern recognition; iterative methods; probability; relaxation theory; computerized pattern recognition; label probabilities; mode detection; multidimensional data clustering; probability density function; relaxation process; Databases; Density functional theory; Electronic switching systems; Entropy; Frequency; Humans; Multidimensional systems; Natural languages; Neodymium; Statistics;
Journal_Title :
Pattern Analysis and Machine Intelligence, IEEE Transactions on
