Session Program


  • 12 July 2017
  • 02:30PM - 04:30PM
  • Room: Santa Lucia
  • Chairs: Gisella Facchinetti and Franco Cutugno

Decision modelling under uncertainty. Applications in Economics and Social Sciences and Social Signal Processing-I

Abstract - The measurement of volatility is of fundamental importance in finance. The standard market practice adopted for the computation of a volatility index imposes to discard some option prices quoted in the market, resulting in a considerable loss of information. To overcome this drawback, we propose to resort to fuzzy regression methods in order to include all the available information and obtain an informative volatility index for the Italian stock market.
Abstract - It has been noted in the literature that firms rarely follow a single theoretical model when designing their compensation policy. This study illustrates how a fuzzy cluster analysis can be helpful in understanding the way employees are rewarded according to firms' specificity and market conditions. For this purpose, we convert linked employer-employee data (LEED) into firm level data prior to fuzzy clustering. Then, we explore the particular distribution of firms on the emerged fuzzy partition to sort them by compensation policy and, eventually, to examine the potential factors behind a specific option.
Abstract - This study proposes a statistical hypothesis test with fuzzy data which is represented in a form of interval data to perform one-way analysis of variance (ANOVA) test with fuzzy data. We adopt 2015 LEGO consolidated income as an empirical study for analyzing the main factors that affecting the LEGO annual profit by using one-way ANOVA Test with interval data. The results shown that each impact factor has different influence levels by means of testing their means. To confirm the most influential impact factor of 2015 LEGO annual profit, we provided the pair-wise comparisons in the evaluation. The comparison results shown that the greater financial income is decided from revenue factor and the most risk will be revealed from expenses factor. We concluded that the proposed method is able to assist an enterprise in identifying the main impact factors which are crucial for financial decision making.
Abstract - A non-traditional approach about the measurement of agents' preference stability is introduced. This contribution focus on measuring preference consensus at different moments under the assumption of considering the following evaluations: approved, undecided and disapproved. To this aim, the concept of preference stability measure is defined as well as a particular one, the sequential preference stability measure, taking into account any two successive time moments. Finally and in order to highlight the good behaviour of novel measures, some properties are also provided.
Abstract - In this paper we propose a parametric way to associate to an interval-valued fuzzy set its evaluation useful for its ranking. The novelty of this paper is connected with the fact that we follow a line based on its alpha-cuts and the parametric formulation we obtain, leaves to the decision maker a wide freedom. For particular values of these parameters we obtain Nie and Tan defuzzification method that, in its classical definition, shows only the evaluation, but looking at it in this new version we obtain further information. The proposed methodology is then applied to risk profiling of a bank client using an interval type-2 fuzzy logic system.
Abstract - A dynamic portfolio allocation is discussed in asset management with fuzziness. By perception-based extension for fuzzy random variables, a dynamic portfolio model for value-at-risks of fuzzy random variables is introduced. By dynamic programming and mathematical programming, this paper derives analytical solutions for the optimization problem. A numerical example is given to demonstrate the results.