Title :
An improved stochastic gradient algorithm for principal component analysis and subspace tracking
Author :
Dehaene, Jeroen ; Moonen, Marc ; Vandewalle, Joos
Author_Institution :
Belgian Nat. Fund for Sci. Res., Katholieke Univ., Leuven, Belgium
fDate :
10/1/1997 12:00:00 AM
Abstract :
We propose a new stochastic gradient algorithm for principal component analysis and subspace tracking, requiring O(nm) operations per update, where n is the number of input signals, and m is the signal subspace dimension. A parallel version with problem size independent throughput is obtained at the expense of O(n2) additional flops
Keywords :
computational complexity; parallel algorithms; signal flow graphs; signal processing; stochastic processes; tracking; computational complexity; input signals; parallel algorithm; principal component analysis; problem size independent throughput; signal flow graph; signal processing; signal subspace dimension; stochastic gradient algorithm; subspace tracking; Adaptive filters; Adaptive signal processing; Eigenvalues and eigenfunctions; Flow graphs; Parallel algorithms; Principal component analysis; Signal processing algorithms; Singular value decomposition; Stochastic processes; Throughput;
Journal_Title :
Signal Processing, IEEE Transactions on
