DocumentCode :
2924047
Title :
Topic graph based transfer learning via generalized KL divergence based NMF
Author :
Kimura, Keigo ; Yoshida, Tetsuya
Author_Institution :
Grad. Sch. of Inf. Sci. & Technol., Hokkaido Univ., Sapporo, Japan
fYear :
2011
fDate :
8-10 Nov. 2011
Firstpage :
330
Lastpage :
335
Abstract :
We propose a topic graph based transfer learning method based on Non-negative Matrix Factorization (NMF) with generalized Kullback-Leibler (KL) divergence. Based on the Frobenius norm based NMF, a transfer learning method was proposed based on the similarity of feature spaces. We extend the previous method by utilizing generalized KL divergence based NMF so that better probabilistic interpretation can be obtained with the divergence. The proposed method is formalized as the minimization of an objective function under the divergence, and an auxiliary function for the objective function is defined. From the auxiliary function, we derive a learning algorithm with multiplicative update rules, which are guaranteed to converge. The proposed method is evaluated in terms of document clustering over several well-known benchmark datasets. Extensive experiments have been conducted on the datasets, and comparison with other transfer learning methods as well as state-of-the-art NMF methods is reported.
Keywords :
document handling; graph theory; learning (artificial intelligence); matrix decomposition; pattern clustering; Frobenius norm based NMF; auxiliary function; document clustering; generalized KL divergence based NMF; generalized Kullback-Leibler divergence; multiplicative update rules; nonnegative matrix factorization; probabilistic interpretation; topic graph based transfer learning method; Clustering algorithms; Degradation; Laplace equations; Learning systems; Minimization; Probabilistic logic; Vectors; Kullback-Leibler divergence; Non-negative Matrix Factorization; Sparse Coding; Transfer Learning;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Granular Computing (GrC), 2011 IEEE International Conference on
Conference_Location :
Kaohsiung
Print_ISBN :
978-1-4577-0372-0
Type :
conf
DOI :
10.1109/GRC.2011.6122617
Filename :
6122617
Link To Document :
بازگشت