文档介绍：890 IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE, VOL. 23, NO. 8, AUGUST 2001
Short Papers___________________________________________________________________________________________________
ic Algorithm Wavelet Design paper and the classification performance of the GA-designed
wavelets pared to that of classical (Cohen et al. [6] wavelets,
for Signal Classification as well as to wavelets designed by minimizing the error in the
wavelet representation [4]. It is important to note that the cost
Eric Jones, Member, IEEE, function employed in [4] allowed a direct solution, without the
Paul Runkle, Member, IEEE, need for a GA. However, that cost function does not permit one to
Nilanjan Dasgupta, Student Member, address the ultimate goal of improved classification. The cost
IEEE, Luise Couchman, and function introduced here is based explicitly on classifier perfor-
Lawrence Carin, Fellow, IEEE mance, this not permitting a direct solution for the optimal
wavelets. Therefore, we have employed this classification-based
cost function in a GA.
AbstractÐBiorthogonal wavelets are applied to parse multiaspect transient
scattering data in the context of signal classification. A language-based ic
algorithm is used to design wavelet filters that enhance classification performance. 2LIFTING-BASED WAVELET DESIGN
The biorthogonal wavelets are implemented