Search results

This paper describes the simulation of a nearest neighbour searching algorithm for document retrieval using a pool of microprocessors. The documents in a database are organised in a multi‐dimensional binary search tree, and the algorithm identifies the nearest neighbour for a query by a backtracking search of this tree. Three techniques are described which allow parallel searching of the tree. A PASCAL‐based, general purpose simulation system is used to simulate these techniques, using a pool of Transputer‐like microprocessors with three standard document test collections. The degree of speed‐up and processor utilisation obtained is shown to be strongly dependent upon the characteristics of the documents and queries used. The results support the use of pooled microprocessor systems for searching applications in information retrieval.

This article reviews the state of the art in automatic indexing, that is, automatic techniques for analysing and characterising documents, for manipulating their descriptions in searching, and for generating the index language used for these purposes. It concentrates on the literature from 1968 to 1973. Section I defines the topic and its context. Sections II and III consider work in syntax and semantics respectively in detail. Section IV comments on ‘indirect’ indexing. Section V briefly surveys operating mechanized systems. In Section VI major experiments in automatic indexing are reviewed, and Section VII attempts an overall conclusion on the current state of automatic indexing techniques.

Some parameters and techniques in use for describing the results of tests on IR systems are analysed. Several considerations outside the scope of the usual 2 x 2 table are relevant to the choice of parameters. In particular, a variable which produces a ‘performance curve’ of a system corresponds to an extension of the 2 x 2 table. Also, the statistical relationships between parameters are all‐important. It is considered that precision is not such a useful measure of performance (in conjunction with recall) as fallout. A more powerful alternative to Cleverdon's ‘inevitable inverse relationship between recall and precision’ is proposed and justified, namely that the recall‐fallout graph is convex.

Reports a laboratory comparison of the effectiveness and efficiency of five index languages in the subject area of library and information science; three post‐co‐ordinate languages, Compressed Term, Uncontrolled, and Hierarchically Structured, and two pre‐co‐ordinate ones, Hierarchically Structured and Relational Indexing. Eight test comparisons were made, and factors studied were index language specificity and linkage, indexing specificity and exhaustivity, method of co‐ordination, the precision devices of partitioning and relational operators, and the provision of context in the search file. Full details of the test and retrieval results are presented.

After considering the search process and functions of index entries, a classification of entry types is offered, based on index term context, predominant term order, and between‐term function words. Then a multiple entry generation scheme is described, comprising rules for term manipulation, input and output. After discussing access points and cross reference measures, a preliminary linguistic analysis is given, showing links with psycholinguistics. The study forms the basis of a current laboratory investigation (EPSILON) into a number of entry types.

This article reports an attempt to understand how a new non‐Boolean ranked output online search facility works. In December 1993, Data‐Star Dialog released a relevance ranking tool known as TARGET (Dialog 1993b). There is considerable research activity into text retrieval using ranking methods, as instanced by the TREC experiments (Harman 1993), and professional online searchers may wish to know exactly what the ranking algorithm does in order to be able to exploit the facility to best advantage. Though it has not been possible to ‘crack’ the algorithm to the level of calculating its match to the nearest per cent, it is possible to see three or four factors at work in the way TARGET produces items ranked in decreasing order of match with a query. It is emphasised that the analysis presented here is based only on records from one bibliographic database. A more extensive and comparative study of bibliographic and full‐text databases would be needed to provide universal and definitive findings.

This paper describes the development of a genetic algorithm (GA) for the assignment of weights to query terms in a ranked‐output document retrieval system. The GA involves a fitness function that is based on full relevance information, and the rankings resulting from the use of these weights are compared with the Robertson‐Sparck Jones F4 retrospective relevance weight. Extended experiments with seven document test collections show that the ga can often find weights that are slightly superior to those produced by the deterministic weighting scheme. That said, there are many cases where the two approaches give the same results, and a few cases where the F4 weights are superior to the ga weights. Since the ga has been designed to identify weights yielding the best possible level of retrospective performance, these results indicate that the F4 weights provide an excellent and practicable alternative. Evidence is presented to suggest that negative weights may play an important role in retrospective relevance weighting.

Engagement, motivation, and persistence are usually associated with positive outcomes. However, too much of it can overtax our psychophysiological system and put it at risk. On the basis of a neuro-dynamic personality and self-regulation model, we explain the neurobehavioral mechanisms presumably underlying engagement and how engagement, when overtaxing the individual, becomes automatically inhibited for reasons of protection. We explain how different intensities and patterns of engagement may relate to personality traits such as Self-directedness, Conscientiousness, Drive for Reward, and Absorption, which we conceive of as functions or strategies of adaptive neurobehavioral systems. We describe how protective inhibitions and personality traits contribute to phenomena such as disengagement and increased effort-sense in chronic fatigue conditions, which often affect professions involving high socio-emotional interactions. By doing so we adduce evidence on hemispheric asymmetry of motivation, neuromodulation by dopamine, self-determination, task engagement, and physiological disengagement. Not least, we discuss educational implications of our model.

Term position information, as provided in some Boolean systems in the form of field restriction and term proximity, is reviewed and its value assessed. Non‐Boolean retrieval in the form of the ranked output experiment has not so far used term position information but has concentrated on schemes of term weighting. The use of term proximity devices is proposed here by analogy with Boolean techniques and seven algorithms are devised to incorporate the ideas of sentence matching, proximate terms, term order specification and term distance computations. It is hypothesised that term position will act as a precision device. A new search experiment is then described in which a test collection is processed into sentences and then output ranking using term position is obtained. Results are given for five algorithms compared against quorum searching as the benchmark. The best result increased the precision ratio by 18% and used proximate matching term pairs in sentences plus a distance component.

Present and possible future developments in the techniques of document management are reviewed, the major ones being text retrieval and scanning and OCR. Acquisition, indexing and thesauri, publishing and dissemination and the document management industry are also addressed. The emerging standards are reviewed and the impact of the Internet is analysed.