Distributed Agents to Retrieve Web Intelligence
  
email E-mail    
HOME                 Registration        About us  Alliances  Contact us |
 Attractions    Bibliography    Classified Papers    Search    Infomaps    Site Map    i-Web    Images Bank    White Papers  
Login Area 11111
 User: 
 Password:
Forgot your password?
Register





free counters
Agents Tour | AI-Lab | Darwin Tour | Faq | News | Newsletter | Pag's | Press Releases | Search Tutorial

Darwin-FIRST Faq

How the HKM evolves?
a) i-URL’s content Once the HKM first version is created starts the evolution process. The HKM degrades with time because users’ disinterest, content obsolescence, sites migration, and sites death. The HKM upgrades via users’ interactions, associated press release centers, Websites’ news and announcement centers, submissions, etc. There are two kinds of upgrades, new URL’s needed to satisfy a given keyword, inexistent in the actual Thesaurus, and new suggested URL’s. The degradation is compensated via retrieval agents, retrievebots that navigate the Web searching for a set of Website candidates for each death i-URL. The upgrade is performed by complex procurebot agents that provide sets of Website candidates as well. Ideally map up and downs could be performed automatically because agents are tuned up to provide candidates sorted by rank, the top the better. However we do not recommend this procedure, at least initially, leaving the final decision to the Chief Editor, the only authorized person to interrupt and to interfere via the Desktop with FIRST. The risk is to become dirty the map because eventual agents failure to detect wrong URL’s. Anyhow, for each selected URL a human expert edits the corresponding i-URL. Notwithstanding, once the HKM initial version is built, the evolution task is not too much burden. Effectively, there are “stable” and low volatile Major Subjects like for instance Philosophy, Religion, Mathematics, highly unstable and extremely volatile like for instance News, and intermediate like for instance computing. In the average we estimate from 100 to 150 daily new i-URL’s for the whole map.

b) Thesaurus evolution Together and in correspondence with i-URL’s ups and downs there are Thesaurus ups and downs. Normally each keyword is sustained by more than one i-URL but it could sometimes happen that the disappearance of a keyword gives place to a special retrieving procurebot mission. Anyhow when new keywords appear at least one i-URL will be uploaded feeding the Thesaurus with the new keyword and some others that could be already existent or not.

c) Logical Tree evolution As we will see in the search process question, the query has the form of a pair (k, s), keyword, subject, that is a specific keyword belonging to a specific subject. Then as time passes by any subject and sub subjects from root to leaves will have a relative popularity weight that will serve at large as a metric of our actual Logical Tree coherence. Special structure agents, structurebots, will suggest Logical Tree changes to the Chief Editor, for instance the deletion of a branch or the upgrade of a leaf. Besides that, the Potential Affinity Groups, PAG’s, from the users’ side, will suggest new subjects as well.


Back
Back to the last page


Copyright © 2003-2013 Darwin! Inc. All rights reserved.