British Library R & D

Report no. 5587

NEW MODELS IN PROBABILISTIC INFORMATION RETRIEVAL

C.J. van Rijsbergen S.E. Robertson M.F. Porter

Computer Laboratory University of Cambridge Corn Exchange Street Cambridge CB2 3QG England

(c) 1980

CONTENTS

Introduction 1 2 3 4 5 6 7 Establishing the NPL test collection at Cambridge The selection of good search terms Probabilistic models of indexing and searching Loose ends Programs for setting up document test collections An algorithm for suffix stripping The CAMIR system