Interactive and nonparametric modeling of preferences on an ordinal scale using small data

Download
2014
Erişkin, Levent
In this study, we consider learning preference structure of a Decision Maker (DM). Many preference modeling problems in a variety of fields such as marketing, quality control and economics, involve possibly interacting criteria, and an ordinal scale is used to express preference of objects. In these cases, typically underlying preference structure of the DM and distribution of criteria values are not known, and only a few data can be collected about the preferences of the DM. For developing a preference model under such circumstances, we propose using nonparametric Statistical Learning approaches interactively. In particular, we employ Active Learning by asking a preference question to the DM at each step and try to reach a close approximation to the correct model in a small number of steps. Our experimental analysis proves that the proposed approach outperforms a “naive” approach where subsequent questions are asked randomly. In the study, we also provide algorithmic recommendations for modeling different underlying value functions, if information is available about the form of the preference structure and/or distribution of criteria values. This study can be regarded as a pioneering approach considering that Statistical Learning based approaches in the literature have been developed and tested based on a relatively large preference information and they do not interact with the DM in model developing process while Multi Criteria Decision Aid based approaches typically ignore interactions among the criteria, suffer from generalization ability, and have no concern about predicting equally good everywhere in the criteria domain.

Suggestions

Interactive and nonparametric modeling of preferences on an ordinal scale using small data
Erişkin, Levent; Köksal, Gülser (2016-12-15)
In this study, we consider learning preference structure of a Decision Maker (DM). Many preference modeling problems in a variety of fields such as marketing, quality control and economics involve possibly interacting criteria, and an ordinal scale is used to express preference of objects. In these cases, typically underlying preference structure of the DM and distribution of criteria values are not known, and only a few data can be collected about the preferences of the DM.
An interactive ranking-based multi-criteria choice algorithm with filtering: Applications to university selection
Karakaya, Gülşah (Orta Doğu Teknik Üniversitesi (Ankara, Turkey), 2019-6)
In this study, we develop an interactive algorithm to converge to the most preferred alternative of a decision maker (DM) among a set of discrete alternatives. The algorithm presents a limited number of alternatives to the DM and collects preference ranking of them iteratively. The preferences are modeled by a flexible and realistic preference function. To improve the performance, the alternatives presented are determined by a filtering method. We compare our algorithm with benchmark algorithms on nume...
Interactive approaches for biobjective problems with progressively changing solution sets
Karakaya, Gülşah (Wiley, 2021-01-01)
In this study, we develop interactive approaches to find a satisfactory alternative of a decision maker (DM) having a quasiconvex preference function where the alternative set changes progressively. In this environment, we keep searching the available set of alternatives and estimating the preference function of the DM. As new alternatives emerge, we make better use of the available preference information and eventually converge to a preferred alternative of the DM. We test our approaches on biobjective, mu...
Interactive Approaches to Multiple Criteria Sorting Problems: Entropy-Based Question Selection Methods
Özarslan, Ali; Karakaya, Gülşah (2022-08-01)
In this study, interactive approaches for sorting alternatives evaluated on multiple criteria are developed. The possible category ranges of alternatives are defined by mathematical models iteratively under the assumption that the preferences of the decision maker (DM) are consistent with an additive utility function. Simulation-based and model-based parameter generation methods are proposed to hypothetically assign the alternatives to categories. A practical approach to solve the incompatibility problem of...
A Probabilistic and interactive approach to multiple criteria sorting
Mutlu, Sinem; Köksalan, Murat; Serin, Yaşar Yasemin; Department of Industrial Engineering (2015)
In this thesis, we develop a method in order to assign alternatives that are evaluated by multiple criteria into preference ordered classes probabilistically. Our motivation is that; when there are large sets of alternatives, placement could be realized in a fast and effective way based on a reasonable misclassification ratio. We assume that the underlying utility function of Decision Maker (DM) is additive. We first ask DM to place reference alternatives into the classes. We develop an interactive probabil...
Citation Formats
L. Erişkin, “Interactive and nonparametric modeling of preferences on an ordinal scale using small data,” Ph.D. - Doctoral Program, Middle East Technical University, 2014.