Title :
A Fast Iterative Algorithm for Identifying Feature Scales and Signed Fuzzy Measures in Generalized Choquet Integrals
Author :
Deng, Xutao ; Wang, Zhenyuan
Author_Institution :
Coll. of Inf. Sci. & Technol., Nebraska Univ., Omaha, NE
Abstract :
We develop an iterative algorithm for identifying approximate optimal scales for feature attributes of generalized Choquet integrals. Based on the scales, the optimal values of fuzzy measures are also obtained immediately using least square method. The new algorithm is significantly faster than previous approaches using evolutionary computing. The experimental results produced by the iterative algorithm are also better than those from genetic algorithms
Keywords :
fuzzy set theory; integral equations; iterative methods; data mining; decision making; evolutionary computing; feature scales; generalized Choquet integrals; genetic algorithms; iterative algorithm; least square method; signed fuzzy measures; Data mining; Educational institutions; Genetic algorithms; Information science; Iterative algorithms; Least squares approximation; Least squares methods; Mathematics; Parameter estimation; Predictive models;
Conference_Titel :
Fuzzy Systems, 2005. FUZZ '05. The 14th IEEE International Conference on
Conference_Location :
Reno, NV
Print_ISBN :
0-7803-9159-4
DOI :
10.1109/FUZZY.2005.1452373
