Showing papers on "Inverted index published in 1979"

Searching MEDLINE in English: a prototype user interface with natural language query, ranked output, and relevance feedback

Abstract: Developing an English sentence (natural language) query capability for today's operational on-line retrieval systems promises to have a profound impact on bibliographic retrieval. Although much research has been done over the past two decades, experimental systems have not been implemented in a large operational environment. In a prototype system called CITE (Current Information Transfer in English), the authors have demonstrated a technical solution for implementing an operational natural language interface to MEDLINE, the most heavily used of the National Library of Medicine's 19 data bases. Important in the design of the system were the functional considerations of identification of search terms, combinatorial searching, weighting and ranked output, relevance feedback, and automatic query modification. Technical considerations focused on maintenance of quick response time and internal efficiency. Of critical importance was the solution of the technical problems of merging postings lists to achieve the logical equivalent of combinatorial searching and computing document ranking weights. The principles employed in the system design are of a general nature and can be transferred to any inverted file system. 2 figures.

Design of a balanced multiple-valued file-organization scheme with the least redundancy

Abstract: A new balanced file-organization scheme of order two for multiple-valued records is presented. This scheme is called HUBMFS2 (Hiroshima University Balanced Multiple-valued File-organization Scheme of order two). It is assumed that records are characterized by m attributes having n possible values each, and the query set consists of queries which specify values of two attributes. It is shown that the redundancy of the bucket (the probability of storing a record in the bucket) is minimized if and only if the structure of the bucket is a partite-claw. A necessary and sufficient condition for the existence of an HUBMFS2, which is composed exclusively of partite-claw buckets, is given. A construction algorithm is also given. The proposed HUBMFS2 is superior to existing BMFS2 (Balanced Multiple-valued File-organization Schemes of order two) in that it has the least redundancy among all possible BMFS2's having the same parameters and that it can be constructed for a less restrictive set of parameters.

A Design for a List Merging Network

Abstract: In inverted file database systems (and in a number of other applications), much of the processing consists of merging two or more previously sorted lists based on a Boolean expression. Conventional general purpose computers are not well suited for this type of operation, with even well written programs utilizing less than 10 percent of the available memory capacity for actual processing of data. However, a low-cost special purpose backend processor can be used to efficiently handle the list merging operations, freeing the conventional processor for other uses.

Inverted File Organization in the Information Retrieval System Based on Thesaurus with Weights.

Abstract: The inverted file structure is often used to organize data in the information retrieval system. When the hierarchy relation on the set descriptors and weights of descriptors in document description would be taken into account, the conventional concept of the inverted file may be extended. Properties and operations on inverted files, which are used in system based on thesaurus with weights, are studied in this paper.

The use of normal multiplication tables for information storage and retrieval

Abstract: This paper describes a method for the organization and retrieval of attribute based information systems, using the normal multiplication table as a directory for the information system. Algorithms for the organization and retrieval of information are described. This method is particularly suitable for queries requesting a group of information items, all of which possess a particular set of attributes (and possibly some other attributes as well). Several examples are given; the results with respect to the number of disk accesses and disk space are compared to other common approaches. Algorithms evaluating the appropriateness of the above approach to a given information system are described. For a certain class of information systems, the normal multiplication table method yields far more rapid retrieval with a more economical space requirement than conventional systems. Moreover this method incorporates an improved modification of the inverted file technique.