AbstractIn this report, we are interested in the determination of the number of clusters for symbolic data described by interval, multi-valued and modal variables and by a combination of this three type. We adapt the five best methods of determination of the number of clusters stemmed from the study of Milligan and Cooper to the program of symbolic classification Sclust as to four hierarchical methods of classification (single linkage, complete linkage, Ward and centroid). We compare the distance available in DISS module of SODAS software and with the more classical distances (L1, L2, Hausdorff et de De Carvalho). We test these methods on various artificial and real data sets and analyse the obtained results.
|Date of Award||2004|
|Supervisor||Andre Hardy (Supervisor), Jean Paul Rasson (Jury) & Pascale Lallemand (Jury)|
Méthodes de détermination du nombre de classes pour des objets symboliques
Troclet, J. (Author). 2004
Student thesis: Master types › Master in Mathematics