Title :
Sparse principal component of a rank-deficient matrix
Author :
Asteris, Megasthenis ; Papailiopoulos, Dimitris S. ; Karystinos, George N.
Author_Institution :
Dept. of Electr. Eng., Univ. of Southern California, Los Angeles, CA, USA
fDate :
July 31 2011-Aug. 5 2011
Abstract :
We consider the problem of identifying the sparse principal component of a rank-deficient matrix. We introduce auxiliary spherical variables and prove that there exists a set of candidate index-sets (that is, sets of indices to the nonzero elements of the vector argument) whose size is polynomially bounded, in terms of rank, and contains the optimal index-set, i.e. the index-set of the nonzero elements of the optimal solution. Finally, we develop an algorithm that computes the optimal sparse principal component in polynomial time for any sparsity degree.
Keywords :
principal component analysis; sparse matrices; auxiliary spherical variables; candidate index-sets; optimal index-set; optimal sparse principal component; polynomial time; rank-deficient matrix; sparsity degree; Complexity theory; Matrix decomposition; Optimization; Polynomials; Principal component analysis; Sorting; Sparse matrices;
Conference_Titel :
Information Theory Proceedings (ISIT), 2011 IEEE International Symposium on
Conference_Location :
St. Petersburg
Print_ISBN :
978-1-4577-0596-0
Electronic_ISBN :
2157-8095
DOI :
10.1109/ISIT.2011.6034216