Issue No. 06 - November/December (2009 vol. 15)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/TVCG.2009.140
Yanhua Chen , Wayne State University, Detroit, MI
Lijun Wang , Wayne State University, Detroit, MI
Ming Dong , Wayne State University, Detroit, MI
Jing Hua , Wayne State University, Detroit, MI
With the rapid growth of the World Wide Web and electronic information services,text corpus is becoming available on-line at an incredible rate.By displaying text data in a logical layout (e.g., color graphs),text visualization presents a direct way to observe the documentsas well as understand the relationship between them.In this paper, we propose a novel technique, Exemplar-based Visualization (EV), to visualizean extremely large text corpus. Capitalizing on recent advances in matrixapproximation and decomposition, EV presents a probabilistic multidimensional projection modelin the low-rank text subspace with a sound objective function. The probability of each document proportion to the topics is obtained through iterative optimization andembedded to a low dimensional space using parameter embedding.By selecting the representative exemplars, we obtain a compactapproximation of the data. This makes the visualization highly efficient and flexible. In addition, the selected exemplars neatly summarize the entire data set and greatly reduce the cognitiveoverload in the visualization, leading to an easier interpretation oflarge text corpus. Empirically, we demonstrate the superior performance of EVthrough extensive experiments performed on the publicly available text data sets.
Exemplar, large-scale document visualization, multidimensional projection.
Y. Chen, J. Hua, M. Dong and L. Wang, "Exemplar-based Visualization of Large Document Corpus (InfoVis2009-1115)," in IEEE Transactions on Visualization & Computer Graphics, vol. 15, no. , pp. 1161-1168, 2009.