Eigenvector space model to capture features of documents



Eigenvectors are a special set of vectors associated with a linear system of equations. Because of the special property of eigenvector, it has been used a lot for computer vision area. When the eigenvector is applied to information retrieval field, it is possible to obtain properties of documents data corpus. To capture properties of given documents, this paper conducted simple experiments to prove the eigenvector is also possible to use in document analysis. For the experiment, we use short abstract document of Wikipedia provided by DBpedia as a document corpus. To build an original square matrix, the most popular method named tf-idf measurement will be used. After calculating the eigenvectors of original matrix, each vector will be plotted into 3D graph to find what the eigenvector means in document processing. 


eigenvector, Vector Space Model, Natural Language Processing, document analyzing, Information Retrieval, text mining

Full Text:



William S. Nobel, What is a support vector machine?, 1565-1567 (Nature Biotechnology, 2006), 24.

Thomas K. Landauer, Peter W. Foltz, and Darrell Lahm, An Introduction to Latent Semantic Analysis, 259-284 (Discourse Processes, 1998), 25.

Matthew A. Turk and Alex P. Pentland, Face recognition using eigenfaces, 586-591 (IEEE Comput. Sco. Press, 1991), 3.

Amy N. Langville and Carl D. Meyer, A Survey of Eigenvector Methods for Web Information Retrieval, 135-161, (SIAM, 2005), 47.

Chang-Beom Lee, Min-Soo Kim, Ki-Ho Lee, Guee-Sang Lee, and Hyuk-Ro Park, Document Thematic words Extraction Using Principal Component Analysis, 747-754 (KIISE, 2002), 29.

Copyright (c) 2011 author

Creative Commons License
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.

In order to comply with GDPR, this site does not allow free registration. Please contact us at: ashues@spiruharet.ro

GDPR Policy:

Please read the statement below:
Processing of personal data and free movement of these data

Registering with this site the author agree with the collection, processing and use of his personal data, exclusively within the ASHUES journal.




To crown and encourage research efforts of the authors, at the end of a year of publishing our journal board will award online the best papers by distinctions:

1. Best Original Paper Award -  for the paper that has brought something completely new in the reader's attention (a new concept, a new trend, a new proposal in research, etc.)

2. Excellence Award - for the most cited paper and visualized in the online environment during the year

3. First, Second and Third Award- for the best documented and substantiated papers during the year

4. Special Award - to award PhD students and postdocs for the most well documented and substantiated paper