Machine learning and vectorial matching for an image retrieval model: EXPRIM and the system RIVAGE
SIGIR '90 Proceedings of the 13th annual international ACM SIGIR conference on Research and development in information retrieval
Fast and effective query refinement
Proceedings of the 20th annual international ACM SIGIR conference on Research and development in information retrieval
Extended Boolean information retrieval
Communications of the ACM
A blueprint for automatic Boolean query processing
ACM SIGIR Forum
Hi-index | 0.02 |
The well-known relevance feedback process uses information extracted from previously retrieved relevant documents to generate improved search formulations for subsequent search iterations. Methods are outlined in this study for the automatic generation of Boolean search statements based on the natural language texts of initially available search requests and of previously retrieved document excerpts identified as relevant by the user population. The search requests are generated both in a conventional Boolean system and in an extended system in which the normal interpretation of the Boolean connectives is relaxed. Experimental output is included which shows that substantial improvements in retrieval effectiveness are obtainable using the automatic relevance feedback methods.