Type of Document Master's Thesis Author Frederic, John Author's Email Address freddy_dx@yahoo.com URN etd-11202008-234124 Title Examination of Initialization Techniques for Nonnegative Matrix Factorization Degree Master of Science Department Mathematics and Statistics Advisory Committee
Advisor Name Title Dr. Marina Arav Committee Chair Dr. Frank Hall Committee Member Dr. Michael Stewart Committee Member Dr. Rachel Belinsky Committee Member Dr. Zhongshan Li Committee Member Keywords
- Nonnegative matrix factorization
- Initialization
- Spherical K- means
- Compression ratio
- Percent error
- Random Acol
- Random C
Date of Defense 2008-11-13 Availability unrestricted Abstract While much research has been done regarding different Nonnegative Matrix Factorization (NMF) algorithms, less time has been spent looking at initialization techniques. In this thesis, four different initializations are considered. Aftera brief discussion of NMF, the four initializations are described and each one is
independently examined, followed by a comparison of the techniques. Next, each initialization's performance is investigated with respect to the changes in the size of the data set. Finally, a method by which smaller data sets may be used to
determine how to treat larger data sets is examined.
Files
Filename Size Approximate Download Time (Hours:Minutes:Seconds)
28.8 Modem 56K Modem ISDN (64 Kb) ISDN (128 Kb) Higher-speed Access frederic_john_t_200812_ms.pdf 734.67 Kb 00:03:24 00:01:44 00:01:31 00:00:45 00:00:03