City Research Online

Abstract

Research was conducted into Farradane's system of relational indexing (used to represent the explicit relationship between concepts and thereby express meaning) in order to ascertain whether or not it provided an appropriate indexing language in a selective dissemination of information (SDI) system. Five sections of Metals Abstracts, relating to the properties of metals, were adopted as the data base for an experimental system which was operated for six months. The abstracts were indexed by relational indexing techniques and a card index established. Forty-three volunteer users whose interests lay in this subject field provided interest profiles against which the data base was searched and the users also assessed the abstracts for relevance. The system was evaluated by recall - precision and recall - fallout diagrams, with Q and (Mx P) being used as overall performance measures. Average system performance gave ranges of 0.69 ~ 0.82 for recall, 0.72 - 0.87 for precision and 0.004 - 0.003 for fallout. Various browsing techniques (generic browsing, conceptual browsing, condensations and the proposed related browsing) were used and evaluated. Relational indexing was successfully applied to more detailed indexing in a different subject area to those of previous attempts; two new uses of one relation and more complex spatial displays were adopted. Various aspects of system performance were examined, including the recall - fallout graph, and a recall - generality relationship (G44 - M) = k] was found. The distribution of results was multi-modal, an explanation for this being sought in relation to profile structure and content. A preliminary equation for determining the number of questions for adequate testing is proposed and an error analysis was performed. The more general areas of testing methodology and user needs are discussed in relation to the SDI situation. Further research embodying methodological and human behavioural aspects is considered justifiable.

Publication Type:	Thesis (Doctoral)
Subjects:	Z Bibliography. Library Science. Information Resources Z Bibliography. Library Science. Information Resources > Z665 Library Science. Information Science
Departments:	School of Science & Technology > Department of Computer Science School of Science & Technology > School of Science & Technology Doctoral Theses Doctoral Theses

Export

Downloads