Abstract

An experimental program was designed to test the effect of using various sources of index terms on retrieval effectiveness and costs. The purpose of the experimental program was to provide quantitative data for making a management decision on indexing procedures for an information storage and retrieval system. The tradeoff between input cost and retrieval effectiveness was the primary factor in the decision. The experiment was designed to provide the needed data. The full technical document text was divided into five categories: title; abstract; table of contents and lists of figures and tables; author-assigned keywords; and the body. An experimental data base of technical documents was created. Manual indexing using thesaurus controlled keywords was performed. The index term source by category and the time required for indexing by category were recorded.