Maximum Entropy PDF Design Using Feature Density Constraints: Applications in Signal Processing

作者: Paul M. Baggenstoss

DOI: 10.1109/TSP.2015.2419189

关键词:

摘要: This paper revisits an existing method of constructing high-dimensional probability density functions (PDFs) based on the PDF at output a dimension-reducing feature transformation. We show how to modify so that it can provide with highest entropy among all PDFs generate given low-dimensional PDF. The is completely general and applies arbitrary transformations. chain-rule described for multi-stage calculations typically used in signal processing. Examples are including MFCC auto-regressive features. Experimental verification results using simulated data provided comparison competing generative methods.

参考文章(25)
Paul M. Baggenstoss, On the Equivalence of Hanning-Weighted and Overlapped Analysis Windows Using Different Window Sizes IEEE Signal Processing Letters. ,vol. 19, pp. 27- 30 ,(2012) , 10.1109/LSP.2011.2175727
P. Mermelstein, Distance measures for speech recognition, psychological and instrumental Pattern Recognition and Artificial Intelligence. pp. 374- 388 ,(1976)
Steven L. Salzberg, Alberto Segre, Programs for Machine Learning ,(1994)
Fei Song, Jihong Cai, Maximum entropy modeling with feature selection for text categorization asia information retrieval symposium. pp. 549- 554 ,(2008) , 10.5555/1786374.1786451
M. J. Hunt, C. Lefebvre, Distance measures for speech recognition National Aeronautical Establishment, Aeronautical Note NAE-AN-57, NRC no 30144. ,vol. 91, pp. 29394- ,(1989)
Andrew McCallum, Dayne Freitag, Fernando C. N. Pereira, Maximum Entropy Markov Models for Information Extraction and Segmentation international conference on machine learning. pp. 591- 598 ,(2000)
Paul M. Baggenstoss, Optimal Detection and Classification of Diverse Short-duration Signals ieee international conference on cloud engineering. pp. 534- 539 ,(2014) , 10.1109/IC2E.2014.96
Shinto Eguchi, Osamu Komori, Atsumi Ohara, Duality of Maximum Entropy and Minimum Divergence Entropy. ,vol. 16, pp. 3552- 3572 ,(2014) , 10.3390/E16073552