Title :
A Practical Visualization Method for Multi-Dimensional Web Document Vectors
Author :
Le, Kang ; Jiwu, Jing ; Ji, Xiang
Author_Institution :
State Key Lab. of Inf. Security, Chinese Acad. of Sci., Beijing, China
Abstract :
Presenting high-dimensional vectorized text set on 2D plane facilitates a quick perception of the similarity of their contents. This paper presents a practical visualization method used in Web document mining system to display the analysis result. The approach is derived from a classic algorithm. Because the original algorithm is not practical for our usage, we make a detailed analysis to the algorithm, and contribute a great improvement in order to put it into practice. In addition, this paper shows a comparative experiment acquired by our algorithm. To evaluate the algorithm we also propose a novel measurement which is supported by the positive experiment result.
Keywords :
Internet; data mining; data visualisation; text analysis; 2D plane facilitate; classic algorithm; document mining system; multidimensional Web document vector; practical visualization method; Algorithm design and analysis; Data visualization; Displays; Feature extraction; Information security; Laboratories; Multidimensional systems; Springs; Text mining; Vectors;
Conference_Titel :
E-Business and Information System Security, 2009. EBISS '09. International Conference on
Conference_Location :
Wuhan
Print_ISBN :
978-1-4244-2909-7
Electronic_ISBN :
978-1-4244-2910-3
DOI :
10.1109/EBISS.2009.5137890
