Nonnegative DEDICOM Based On Tensor Decompositions for Social Networks Exploration

Nonnegative three-way Decomposition into Directional COmponents (DEDICOM) has been found many potential application including analysis of social networks with large-scale, massive datasets. Existing algorithms for DEDICOM involve difficult nonlinear optimizations problems with multiple local minima. The most efficient algorithm employs ill-conditioned Hessian of a cost function and the Levenberg-Marquard algorithm with a free parameter which is difficult optimally to tune. Moreover. The existing algorithms are not suitable for very large-scale problems. To alleviate these problems we propose an efficient algorithm which first decomposes three-way data into Tucker-3 or Tucker-2, and next retrieves the desired DEDICOM matrices from the Tucker factors and the core tensor.

[1]  J. Kruskal,et al.  A two-stage procedure incorporating good features of both trilinear and quadrilinear models , 1989 .

[2]  Jimeng Sun,et al.  MultiVis: Content-Based Social Network Exploration through Multi-way Visual Analysis , 2009, SDM.

[3]  H. Sebastian Seung,et al.  Learning the parts of objects by non-negative matrix factorization , 1999, Nature.

[4]  Andrzej Cichocki,et al.  Nonnegative Matrix and Tensor Factorization T , 2007 .

[5]  Bülent Yener,et al.  Unsupervised Multiway Data Analysis: A Literature Survey , 2009, IEEE Transactions on Knowledge and Data Engineering.

[6]  L. Tucker,et al.  Some mathematical notes on three-mode factor analysis , 1966, Psychometrika.

[7]  Tamara G. Kolda,et al.  Temporal Analysis of Social Networks using Three-way DEDICOM , 2006 .

[8]  Tamara G. Kolda,et al.  Temporal Analysis of Semantic Graphs Using ASALSAN , 2007, Seventh IEEE International Conference on Data Mining (ICDM 2007).

[9]  Brett W. Bader.  Advances in Data Analysis using PARAFAC2 and Three-way , 2009 .

[10]  Brett W. Bader,et al.  The TOPHITS Model for Higher-Order Web Link Analysis∗ , 2006 .

[11]  Andrzej Cichocki,et al.  Novel Alternating Least Squares Algorithm for Nonnegative Matrix and Tensor Factorizations , 2010, ICONIP.

[12]  David J. Marchette,et al.  Scan Statistics on Enron Graphs , 2005, Comput. Math. Organ. Theory.

[13]  Tamara G. Kolda,et al.  Higher-order Web link analysis using multilinear algebra , 2005, Fifth IEEE International Conference on Data Mining (ICDM'05).

[14]  Tamara G. Kolda,et al.  Pattern Analysis of Directed Graphs Using DEDICOM: An Application to Enron Email , 2006 .

[15]  Jimeng Sun,et al.  SmallBlue: Social Network Analysis for Expertise Search and Collective Intelligence , 2009, 2009 IEEE 25th International Conference on Data Engineering.