INN: an intelligent negotiating neural network for information systems: A design model
作者:
Highlights:
•
摘要
Research on the use of online catalogs reveals that information retrieval by subject is the most common form, but the least successful. Most problems are attributed to users' search terms, which are either too broad or too specific. We present an Intelligent Negotiating Neural Network (INN) design model for solving this aspect of online catalogs. The network is designed to act as an electronic information specialist capable of learning to negotiate a patron's query and translate it into a true, well formulated statement prior to accessing an online catalog. The network's architecture includes four dimensions: conceptual graphs for queries, inheritance and recognition, knowledge base, and modules. This architecture corresponds to four elements of the traditional query negotiation interview performed by information specialists. These are: query/concept; types of sources/ answer-providing tools; types of literature; and time frame. Three different sessions of the user interface are presented: negotiation of a broadly stated query; negotiation of a narrowly stated query; and negotiation of a query for which concepts are nonexistent in the knowledge base.
论文关键词:
论文评审过程:Received 22 June 1993, Accepted 18 October 1993, Available online 19 July 2002.
论文官网地址:https://doi.org/10.1016/0306-4573(94)90077-9