An Approach For Detection Of Logical Inconsistency
Loading...
Date
2012-02
Authors
Yadav, Rinki
Journal Title
Journal ISSN
Volume Title
Publisher
Universiti Sains Malaysia
Abstract
Online responses to multi-criteria decision making questionnaires often contain
inconsistencies, due to possible inconsistent interpretations of the questions or the erratic
mental states of the respondents. If such inconsistent responses remain undetected and no
appropriate corrective actions are taken, the questionnaire outcomes will be less reliable.
The main research motivation was to minimize such inconsistent responses by detecting the
logically inconsistent responses dynamically. The conventional approaches compute a
consistency ratio to quantitatively specify the level of inconsistent responses. However, in
most online questionnaires, it is too late to highlight the inconsistency level at the end. They
should be flagged out as soon as each inconsistent response surfaces.
In order to resolve such a problem, this research has been conducted to detect
logically inconsistent responses as soon as they arise. Furthermore, a triad comparison
technique based on seven logical rules is proposed. In the course of the research, it was
found that random arrays of questions did not promote efficient detection of inconsistent
responses. As such, the optimal triad formation rules are applied. It was then found that
detection could start as early as after receiving (n-1) responses, among a total of n(n-1)/2
responses. Subsequent to the (n-1)th response, logical inconsistency detection would then be
performed.
The evaluation results show the correctness and effectiveness of the proposed
approach for inconsistency detection on triads (IDOT). The results further show the
necessity of all seven rules in logical inconsistency detection. The comparison of IDOT with
other conventional detection approaches shows the effectiveness of IDOT in terms of
improved consistency ratio, which is reduced to 6.7% from 16.4%. It is believed that IDOT
for specially sequenced questions is necessary to better and early detect inconsistent
responses.
Description
Keywords
An approach for detection , of logical inconsistency