DocumentCode :
3412509
Title :
Implicit softmax transforms for dimensionality reduction
Author :
Tuerk, Andreas
Author_Institution :
Telecommun. Res. Center, Vienna
fYear :
2008
fDate :
March 31 2008-April 4 2008
Firstpage :
1973
Lastpage :
1976
Abstract :
This paper develops implicit softmax transforms (IST) which are dimensionality reducing transforms that are defined implicitly by minimisation of a weighted sum of Kullback-Leib- ler distances (WKL). The parameters of an IST are trained in combination with the parameters of the polynomial exponents of softmax functions. The resulting gradient of the WKL can be efficiently calculated and the computational effort scales well with the size of the training set. The paper compares IST´s to PCA and LDA in classification experiments with two different types of classifiers on three different datasets, two of them from the UCI machine learning repository. It is shown that IST´s outperform PCA and LDA in a majority of the cases. In one case reducing the dimension with an IST even gives an improvement over the high dimensional baseline system.
Keywords :
data reduction; learning (artificial intelligence); minimisation; pattern classification; polynomials; principal component analysis; transforms; Kullback-Leibler distance; LDA; PCA; dimensionality reduction; implicit softmax transform; machine learning; minimisation; pattern classification; polynomial exponent; Independent component analysis; Information theory; Linear discriminant analysis; Machine learning; Multidimensional signal processing; Mutual information; Pattern recognition; Polynomials; Principal component analysis; Tin; Pattern recognition; information theory; multidimensional signal processing;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Acoustics, Speech and Signal Processing, 2008. ICASSP 2008. IEEE International Conference on
Conference_Location :
Las Vegas, NV
ISSN :
1520-6149
Print_ISBN :
978-1-4244-1483-3
Electronic_ISBN :
1520-6149
Type :
conf
DOI :
10.1109/ICASSP.2008.4518024
Filename :
4518024
Link To Document :
بازگشت