Multi-agent approach for community clustering based on individual ontology annotations

  • Authors:
  • Dmitri Soshnikov

  • Affiliations:
  • Dept. of Numerical Mathematics and Programming, Artificial Intelligence Group, Moscow Aviation Institute (Technical University)

  • Venue:
  • AIS-ADM 2005 Proceedings of the 2005 international conference on Autonomous Intelligent Systems: agents and Data Mining
  • Year:
  • 2005

Quantified Score

Hi-index 0.00

Visualization

Abstract

The paper presents peer-to-peer multi-agent framework for community clustering based on contact propagation in the global network of contacts with individual ontology-based descriptions. The framework allows people to meet interested parties by specifying flexible ontology-based requests and/or proposals and relaying them between the nodes of the system along the most ontologically-prominent directions, thus covering wide yet very targeted audience. It also supports ad hoc meeting using mobile devices such as PDAs, SmartPhones, etc. The system works over standard means of communication such as e-mail/ICQ, and provides user with personalized relaying assistant that can automatically process ontologically-annotated messages, relaying and or responding to them. It can also collect user's knowledge about his contact base, thus acting as a personalized assistant in the certain area of communication.