PUB TYPE Reports- Research/Technical (143)-- Tests /Evaluation Instruments (160)

This project examines anaphora (the linguistic device of abbreviated subsequent reference to a concept) in information retrieval (IR) systems in order to develop procedures to recognize anaphors in text and distinguish between anaphoric and non-anaphoric uses of a given term, estimate the number of anaphors appearing in bibliographic records, and assess the effect on retrieval performance when anaphors are replaced by their referents. In the first phase of the study, rules were developed to form the basis for an automatic procedure to recognize anaphoric terms in bibliographic databases. An examination of the titles and abstracts of 600 documents revealed that only 3.67 true anaphors occurred in the average abstract, suggesting that the effect of treating these terms in some way to improve retrieval performance might be slight. In the second phase, 12 term weighting schemes were used to determine the relevance o