A Novel Bivariate Elliptic Fuzzy Membership Function: A Modeling and Decision-Making Tool for Bike Sharing


BAŞARAN A. A., Basaran M. A., Demir M. O.

Intelligent and Fuzzy Systems - Intelligence and Sustainable Future Proceedings of the INFUS 2023 Conference, İstanbul, Turkey, 22 - 24 August 2023, vol.758 LNNS, pp.77-84 identifier

  • Publication Type: Conference Paper / Full Text
  • Volume: 758 LNNS
  • Doi Number: 10.1007/978-3-031-39774-5_10
  • City: İstanbul
  • Country: Turkey
  • Page Numbers: pp.77-84
  • Keywords: Bike-sharing, Chebyshev’s inequality, Decision making, Elliptic fuzzy membership function
  • Hacettepe University Affiliated: Yes

Abstract

In this paper, the data set generated by Bike Sharing System (BSS) has been modeled by a proposed method called fuzzy bivariate elliptic membership function that generates membership values between an independent variable and dependent variable whose functional form follows an ellipse since all numerical variables following a cyclic pattern such as season, month, hour, weather situation and so on. Besides, each membership value corresponding to each independent and dependent variable is used to find an aggregate outcome of a dependent variable based on a new decision-making tool. Therefore, how both weather and time combinations have an impact on the dependent variable could be derived. Since there is no built-in membership function available, the data set is used to construct a data-driven elliptic fuzzy membership function. Thus, the Chebyshev inequality based on the correlated variables is used to determine both the a and b parameters of the elliptic function representing 95% of the whole dataset.