Retrieval Using Plausible Inference
Abstract
The goal of this project was basic research and development in the area of text-based information systems. A new probabilistic model of text retrieval using Bayesian inference nets was proposed and extended. An efficient implementation of this model for large full-text databases resulted in the INQUERY indexing and retrieval engine. Retrieval experiments carried out using INQUERY an a variety of databases have shown that it produces high-quality results compared to other approaches. We have also made significant progress, both conceptually and with prototype implementations, in showing how text retrieval can be integrated with hypertext and database management systems in a single framework. The success of the INQUERY system in ARPA-sponsored evaluations has led to a number of technology transfer projects.
Document Details
- Document Type
- Technical Report
- Publication Date
- Oct 01, 1993
- Accession Number
- ADA274252
Entities
People
- W. Bruce Croft
Organizations
- University of Massachusetts Amherst