DocumentCode :
1248182
Title :
Adaptive Forward-Backward Greedy Algorithm for Learning Sparse Representations
Author :
Zhang, Tong
Author_Institution :
Stat. Dept., Rutgers Univ., Piscataway, NJ, USA
Volume :
57
Issue :
7
fYear :
2011
fDate :
7/1/2011 12:00:00 AM
Firstpage :
4689
Lastpage :
4708
Abstract :
Given a large number of basis functions that can be potentially more than the number of samples, we consider the problem of learning a sparse target function that can be expressed as a linear combination of a small number of these basis functions. We are interested in two closely related themes: · feature selection, or identifying the basis functions with nonzero coefficients; · estimation accuracy, or reconstructing the target function from noisy observations. Two heuristics that are widely used in practice are forward and backward greedy algorithms. First, we show that neither idea is adequate. Second, we propose a novel combination that is based on the forward greedy algorithm but takes backward steps adaptively whenever beneficial. For least squares regression, we develop strong theoretical results for the new procedure showing that it can effectively solve this problem under some assumptions. Experimental results support our theory.
Keywords :
estimation theory; feature extraction; greedy algorithms; learning (artificial intelligence); least squares approximations; regression analysis; signal reconstruction; adaptive forward-backward greedy algorithm; basis function identification; feature selection; least squares regression; noisy observation; sparse representation learning; sparse target function; target function reconstruction; Accuracy; Algorithm design and analysis; Approximation methods; Greedy algorithms; Machine learning; Matching pursuit algorithms; Noise; Estimation theory; feature selection; greedy algorithms; sparse recovery; statistical learning;
fLanguage :
English
Journal_Title :
Information Theory, IEEE Transactions on
Publisher :
ieee
ISSN :
0018-9448
Type :
jour
DOI :
10.1109/TIT.2011.2146690
Filename :
5895111
Link To Document :
بازگشت