Title :
Wombat: a toolbox to build and optimize a fuzzy regression tree
Author :
Guillou, Xavier Le ; Glorennec, Pierre-Yves
Author_Institution :
Dept. of Comput. Sci., INSA de Rennes/IRISA, Rennes
Abstract :
Wombat is a toolbox allowing to build and optimize a fuzzy regression tree from numerical data. The number and the placement of membership functions are optimized in order to increase the information gain, under the user´s control. Wombat can be used from a console or using a graphical interface written in Tcl/Tk. The built tree can be exported, either in a graphical form (postscript, svg...), or in a file which can be used by an independent program in C language. Wombat will be proposed as a contribution to SCILAB
Keywords :
data mining; fuzzy logic; graphical user interfaces; mathematics computing; optimisation; regression analysis; trees (mathematics); C language; Tcl/Tk; Wombat; constrained optimization; data mining; fuzzy logic; fuzzy regression tree; graphical interface; knowledge extraction; membership function optimization; supervised learning; Constraint optimization; Data mining; Decision trees; Fuzzy logic; Fuzzy sets; Page description languages; Regression tree analysis; Supervised learning; Testing; Tree graphs;
Conference_Titel :
Control Applications, 2005. CCA 2005. Proceedings of 2005 IEEE Conference on
Conference_Location :
Toronto, Ont.
Print_ISBN :
0-7803-9354-6
DOI :
10.1109/CCA.2005.1507283
