NEUROSEMANTIC APPROACH TO BUILDING AUTOMATED INFORMATION RETRIEVAL SYSTEMS

Abstract

Currently, one of the actual problems in the formation of the most informative knowledge models of this subject area is the creation of effective information retrieval systems, primarily on the Internet, as the largest repository of information. To extract the necessary information for this subject area from the Internet, processing of a huge number of heterogeneous documents is required. This is the rather complicated task that requires not only automating the process of searching for information, but also ensuring its semantic content in accordance with the current situation in this subject area. To automate the search process and determine the most informative content of the knowledge model of this subject area, a multi-agent intelligent system for searching and selecting information based on neural networks proposed. This system implements a neurosemantic approach to the semantic adaptation of search information to changes in the current situation in this subject area and the corresponding evolution of the knowledge model based on a genetic algorithm. Ref. 9

Keywords

• intelligent multiagent system
• neural network
• genetic algorithm
• knowledge model
• subject area