Title :
Multivariate Data Classification Using the Distribution Mapping Exponent
Author :
Jirina, Marcel ; Jirina, Marcel
Author_Institution :
Inst. of Comput. Sci., Czech Acad. of Sci., Prague
Abstract :
An exponent similar to the correlation dimension is introduced. This exponent is used for probability density estimation in high-dimensional spaces and for classification of multivariate data. It is also shown that this classifier exhibits significantly better behavior (classification accuracy) than other kinds of classifiers.
Keywords :
data handling; pattern classification; probability; correlation dimension; distribution mapping exponent; high-dimensional spaces; multivariate data classification; probability density estimation; Bayesian methods; Biomedical engineering; Computer science; Content addressable storage; Decision trees; Euclidean distance; Extraterrestrial measurements; Nearest neighbor searches; Neural networks; Probability distribution;
Conference_Titel :
Computational Cybernetics, 2006. ICCC 2006. IEEE International Conference on
Conference_Location :
Budapest
Print_ISBN :
1-4244-0071-6
Electronic_ISBN :
1-4244-0072-4
DOI :
10.1109/ICCCYB.2006.305707
