Authors:
Catalina Cocianu
1
;
Luminita State
2
;
Ion Roşca
1
and
Panayiotis Vlamos
3
Affiliations:
1
Academy of Economic Studies, Romania
;
2
University of Pitesti, Romania
;
3
Ionian University, Greece
Keyword(s):
Principal axes, supervised learning, pattern recognition, data mining, classification, skeleton.
Related
Ontology
Subjects/Areas/Topics:
Biometrics and Pattern Recognition
;
Multimedia
;
Multimedia Signal Processing
;
Telecommunications
Abstract:
Large multivariate data sets can prove difficult to comprehend, and hardly allow the observer to figure out the pattern structures, relationships and trends existing in samples and justifies the efforts of finding suitable methods from extracting relevant information from data. In our approach, we consider a probabilistic class model where each class h ∈ H is represented by a probability density function defined on R n ; where n is the dimension of input data and H stands for a given finite set of classes. The classes are learned by the algorithm using the information contained by samples randomly generated from them. The learning process is based on the set of class skeletons, where the class skeleton is represented by the principal axes estimated from data. Basically, for each new sample, the recognition algorithm classifies it in the class whose skeleton is the “nearest” to this example. For each new sample allotted to a class, the class characteristics are re-computed using a fir
st order approximation technique. Experimentally derived conclusions concerning the performance of the new proposed method are reported in the final section of the paper.
(More)