Sunday, November 20, 2011

Reaction: TileBars: Visualization of Term Distribution Information in Full Text Information Access

A traditional query model for searching text documents would results in randomly ordered results. Many of these may be either irrelevant to us, or, we may not be able to prima facie guess the usefulness of information contained within the indicated documents. As improvement over this, some of the search engines provide weighted sets of scores to indicate the occurrences of key words in the resultant documents.
However, an effective visualization model is envisaged in this paper. The suggested Graphical Visualization model enables the user to reduce considerable time and effort to visually weigh the resultant documents suggested to be scanned. This search engine-design captures user priority such as minimum overlap span, hits and distribution over multiple topics. It graphically renders the results of the search in “density-and-length packed” tiles. This enables us to visually weigh the required information in each document.
Doing a literature survey over internet, this model has been further picturized or adapted by Juan C. Dürsteler of Inf@Vis! The digital magazine of where three level user query with user-defined “search-limits and clusters” is exemplified. This results in a number of tiles for each document, which would appear as below:

The upper row indicates the frequency of the word "Information" in each section of the document, the lower row corresponds to the same concept for "Visualization". However, key words distribution across the tiles as suggested in the paper is conspicuously modified in this simplified picturization.
As a prelude to this paper, “Querying, Navigating and Visualizing a Digital Library Catalog” by Aravindan Veerasamy, Shamkant Navathe of Georgia Institute of Technology, can provide us with a simple abstraction to understand the suggested model. Hence, a supplementary reading of this related literature can be useful.