Revista Interamericana de Bibliotecología
Print version ISSN 0120-0976
SUAREZ DE LA TORRE, María Mercedes et al. Analysis, design and implementation of a deliberative agent for defining context extraction in specialized texts. Rev. Interam. Bibliot [online]. 2009, vol.32, n.2, pp. 59-84. ISSN 0120-0976.
This paper presents the results of the first stage comprised in the research project entitled: Multiagent System for Defining Contexts Extraction, based on ontologies for the Semantic Web. This article mainly aims at showing the analysis, design and implementation of a deliberative agent with a supervised learning mechanism, which permits the identification of defining contexts in specialized texts. The GAIA methodology has been selected to apply the agent in order to reach the stated goal. This methodology provides an increasing set of steps, based on agent- based systems as an organizational design process. The process of defining context extraction has been carried out by means of the GATE tool, which allows the detection of regular expressions in documents syntactically and semantically marked. The defining contexts are obtained in a semi-automatic way as a result of the interaction between the deliberative agent and the corpus selected. The semi-automatic extraction works by means of the design of the following interphases: basic application interphase, interacting with the JADE platform interphase, and agent communication interphase. After the design and the implementation of each of the inter phases mentioned above, it has been concluded that the research works dealing with the defining context extraction and linguistic patterns, do not provide the necessary information, from the syntactic and semantic aspect, to make the machine recognize the defining contexts and to develop a more refined search and retrieval, while using a minimum phase amount.
Keywords : Deliberative agent; defining contexts; specialized contexts; linguistic patterns; semiautomatic information extraction.