Constructions of local orthonormal bases for classification and regression, (with R. R. Coifman), Comptes Rendus Acad. Sci. Paris, Serie I , vol.319, pp.191-196, Jul. 1994.


We describe extensions to the "best-basis" method to construct orthonormal bases which either maximize a class separability for signal classification problems or minimize an estimation error for regression problems. These algorithms reduce the dimensionality of these problems by using basis functions which are well localized in time-frequency plane as feature extractors.

Get the full paper: gzipped PS file or PDF file.

Please email me if you have any comments or questions!
Go back to Naoki's Publication Page