Title :
Heterogeneous radial basis function networks
Author :
Wilson, D. Randall ; Martinez, Tony R.
Author_Institution :
Dept. of Comput. Sci., Brigham Young Univ., Provo, UT, USA
Abstract :
Radial basis function (RBF) networks typically use a distance function designed for numeric attributes, such as Euclidean or city-block distance. This paper presents a heterogeneous distance function which is appropriate for applications with symbolic attributes, numeric attributes, or both. Empirical results on 30 data sets indicate that the heterogeneous distance metric yields significantly improved generalization accuracy over Euclidean distance in most cases involving symbolic attributes
Keywords :
feedforward neural nets; generalisation (artificial intelligence); learning (artificial intelligence); Euclidean distance; generalization accuracy; heterogeneous distance function; heterogeneous radial basis function networks; numeric attributes; symbolic attributes; Application software; Backpropagation; Computer science; Euclidean distance; Genetics; Nearest neighbor searches; Neural networks; Programming profession; Radial basis function networks;
Conference_Titel :
Neural Networks, 1996., IEEE International Conference on
Conference_Location :
Washington, DC
Print_ISBN :
0-7803-3210-5
DOI :
10.1109/ICNN.1996.549079
