FORECASTING OF THE FUZZY UNIVARIATE TIME SERIES BY THE OPTIMAL LAGGED REGRESSION STRUCTURE DETERMINED BASED ON THE GENETIC ALGORITHM

Creative Commons License

Eren M.

ECONOMIC COMPUTATION AND ECONOMIC CYBERNETICS STUDIES AND RESEARCH, cilt.52, sa.2, ss.201-215, 2018 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 52 Sayı: 2
  • Basım Tarihi: 2018
  • Doi Numarası: 10.24818/18423264/52.2.18.12
  • Dergi Adı: ECONOMIC COMPUTATION AND ECONOMIC CYBERNETICS STUDIES AND RESEARCH
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Social Sciences Citation Index (SSCI), Scopus
  • Sayfa Sayıları: ss.201-215
  • Anahtar Kelimeler: Genetic Algorithms, LR-Type Fuzzy Numbers, Fuzzy Least Squares Method, Time Series, Forecasting, LINEAR-REGRESSION, LOGISTIC-REGRESSION, MODEL
  • Ondokuz Mayıs Üniversitesi Adresli: Evet

Özet

Estimation obtained through classical regression model reveals the fitting (or prediction) and projection (or forecast) values with a certain error. This situation leads to loss of information and imprecision of data. However, if the imprecise information is converted to fuzzy data rather than single value, an estimation procedure can be obtained in which observation errors are hidden in fuzzy coefficients. Thus, it would be more realistic to make an interval estimate instead of a single value estimate with a certain margin of error. Therefore, in this study, a novel fuzzy least squares method developed for the variables expressed by LR-type fuzzy numbers, based on the optimal classical lagged regression model structure determined by the genetic algorithm, was addressed. a numerical example to explain how the proposed method is applicable was considered.