Browse Prior Art Database

Adaptive Query System

IP.com Disclosure Number: IPCOM000075235D
Original Publication Date: 1971-Aug-01
Included in the Prior Art Database: 2005-Feb-24
Document File: 2 page(s) / 32K

Publishing Venue

IBM

Related People

Gause, DC: AUTHOR [+3]

Abstract

The system accepts user-supplied character strings and compares them with previously stored strings in order to produce the correct answer. Performance is measured in terms of finding the correct answer with the minimum number of misses and improves with increased use.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 87% of the total text.

Page 1 of 2

Adaptive Query System

The system accepts user-supplied character strings and compares them with previously stored strings in order to produce the correct answer. Performance is measured in terms of finding the correct answer with the minimum number of misses and improves with increased use.

The data structure is initialized by supplying two questions on a subject. The strings are compared, and the longest common substring is termed an operand, with the remaining strings called the operators. The answer is supplied, stored and can be subsequently accessed by supplying either question. To permit a rapid check for identical questions, a hashing function is used upon entry of a question. If hashing fails, an operator/operand search is performed; and if either is located, a new answer is required. Otherwise, the question is compared with all strings to measure similarity.

The method by which strings are judged to be similar is based on structural attributes of the strings, rather than a syntactical and semantic analysis. Features of strings are compared as follows: 1) simple correlation of the number of occurrences of symbols, 2) simple correlation of position indices of symbols in a standard string, 3) sum of absolute values of differences between index vectors in 2), 4) comparison of precedence matrices for strings against standard string, and 5) simple correlation between precedence vectors developed in 4).

Coefficients developed above are X's in the decision funct...