Pairwise comparison

About: Pairwise comparison is a research topic. Over the lifetime, 6804 publications have been published within this topic receiving 174081 citations.

Analytic hierarchy prioritization process in the AHP application development: A prioritization operator selection approach

Abstract: In the analytic hierarchy process, prioritization of the reciprocal matrix is a core issue to influence the final decision choice. Various prioritization methods have been proposed, but none of prioritization methods performs better than others in every inconsistent case. To address the prioritation operator selection problem, this paper proposes the analytic hierarchy prioritization process, which is an objective hierarchy model (without using subjective pairwise comparisons) to approximate the real priority vectors with selection of the most appropriate prioritization operator among the various prioritization candidates, for a reciprocal matrix, and on the basis of a list of measurement criteria. Nine important prioritization operators and seven measurement criteria are illustrated in AHPP. Two previous applications are revised and illustrate the validity and usability of the proposed model. The results show that the most appropriate prioritization operator is dependent of the content of the reciprocal matrix and AHPP is an appropriate method to address the prioritization problem to make better decisions.

ELECCALC—an interactive software for modelling the decision maker's preferences

Abstract: In general, it is difficult to articulate the decision-maker's (DM) preference structure, specially in taking into account several criteria. Most synthetical approaches (single synthetical criterion approach, synthetical outranking approach,…) are based on some a priori information. In this paper, we attempt to analyze such a structure through a descriptive approach. The idea is to explain the global preferences revealed by the DM from pairwise comparisons of reference alternatives. A disaggregation — aggregation interactive procedure like in PREFCALC is used in ELECCALC, which enables a DM to assess the parameters of ELECTRE II.

Modeling the Outcomes of Vote-Casting in Actual Elections

Abstract: How often do events of interest to voting theorists occur in actual elections? For example, what is the probability of observing a voting cycle – an outcome in which no candidate beats all other candidates in pairwise comparison by majority rule? When there is a candidate who beats all others in such pairwise comparisons – a Condorcet winner – what is the probability that a voting method chooses this candidate?What is the probability that voters have an incentive to vote strategically – that is, cast their votes in ways that do not reflect their true preferences? Voting theorists have analyzed these questions in great detail, using a variety of statistical models that describe different distributions of candidate rankings.

Multi-Label Classification with Label Constraints

Abstract: We extend the multi-label classication setting with constraints on labels. This leads to two new machine learning tasks: First, the label constraints must be properly integrated into the classication process to improve its performance and second, we can try to automatically derive useful constraints from data. In this paper, we experiment with two constraint-based correction approaches as post-processing step within the ranking by pairwise comparison (RPC)-framework. In addition, association rule learning is considered for the task of label constraints learning. We report on the current status of our work, together with evaluations on synthetic datasets and two real-world datasets.

An Interaction-Based Test Sequence Generation Approach for Testing Web Applications

Abstract: Web applications often use dynamic pages that interact with each other by accessing shared objects, e.g., session objects. Interactions between dynamic pages need to be carefully tested, as they may give rise to subtle faults that cannot be detected by testing individual pages in isolation. Since it is impractical to test all possible interactions, a trade-off must be made between test coverage (in terms of number of interactions covered in the tests) and test effort. In this paper, we present a test sequence generation approach to cover all pairwise interactions, i.e., interactions between any two pages. Intuitively, if a page P could reach another page Ppsila, there must exist a test sequence in which both P and Ppsila are visited in the given order. We report a test sequence generation algorithm and two case studies in which test sequences are generated to achieve pairwise interaction coverage for two Web applications. The empirical results indicate that our approach achieves good code coverage and is effective for detecting interaction faults in the subject applications.