DocumentCode :
3016435
Title :
Binary Matrix Factorization and Consensus Algorithms
Author :
Fu, Yinghua ; Jiang, Nianping ; Sun, Hong
Author_Institution :
Sch. of Opt.-Electr. & Comput. Eng., Univ. of Shanghai for Sci. & Technol., Shanghai, China
fYear :
2010
fDate :
25-27 June 2010
Firstpage :
4563
Lastpage :
4567
Abstract :
In data mining, SVD is a popular method that has been used for compressing high dimensional data. Binary matrix factorization (BMF) is a variant of SVD. There are two methods for binary factorization compression: the iterative heuristic and greedy algorithms. However, both of them are not perfect in applications. The iterative heuristic does not guarantee the convergence in most cases and greedy algorithms can´t fit the need of large-scale matrices factorization. In this paper a new method is used for BMF: consensus algorithms. Consensus algorithms are a brand-new approach to enumerating all the maximal bicliques for a given graph, which is proved to be an NP-complete problem and can give the solution in incremental polynomial time. For some bipartite graphs, the time complexity is polynomial. Experiments show that when the iterative heuristic does not work, consensus algorithm improves far more badly the efficiency than greedy algorithms, and ensures the stability.
Keywords :
computational complexity; data compression; data mining; graph theory; greedy algorithms; iterative methods; singular value decomposition; NP-complete problem; SVD; binary matrix factorization; bipartite graphs; consensus algorithms; data mining; greedy algorithm; high dimensional data compression; incremental polynomial time; iterative heuristic algorithm; maximal bicliques; time complexity; Algorithm design and analysis; Approximation methods; Greedy algorithms; Image coding; Iterative algorithm; Matrix decomposition; Polynomials; Consensus Algorithms MICA BMF SVD; Iterative Heuristic Rank-one Approximation;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Electrical and Control Engineering (ICECE), 2010 International Conference on
Conference_Location :
Wuhan
Print_ISBN :
978-1-4244-6880-5
Type :
conf
DOI :
10.1109/iCECE.2010.1455
Filename :
5631749
Link To Document :
بازگشت