A new approach based on regression analysis and mathematical programming to multi-group classification problems

Creative Commons License

DOĞAN M. İ., Orman A., ÖRKCÜ M., ÖRKCÜ H. H.

JOURNAL OF THE FACULTY OF ENGINEERING AND ARCHITECTURE OF GAZI UNIVERSITY, cilt.34, sa.4, ss.1939-1955, 2019 (SCI-Expanded) identifier identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 34 Sayı: 4
  • Basım Tarihi: 2019
  • Doi Numarası: 10.17341/gazimmfd.571643
  • Dergi Adı: JOURNAL OF THE FACULTY OF ENGINEERING AND ARCHITECTURE OF GAZI UNIVERSITY
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, TR DİZİN (ULAKBİM)
  • Sayfa Sayıları: ss.1939-1955
  • Anahtar Kelimeler: Classification problem, mathematical programming, regression analysis, two-stage approach, DEA-DISCRIMINANT ANALYSIS, OPTIMAL CRITERION WEIGHTS, MINIMIZING DEVIATIONS, DECISION TREE, MODEL, OPTIMIZATION, FISHER, FORMULATIONS, CLASSIFIERS, ALGORITHM
  • Gazi Üniversitesi Adresli: Evet

Özet

In this study, for solving multi-group classification problems, a new two-stage hybrid classification method based on regression analysis and mathematical programming has been developed. In the first step of the proposed method, the classification score of each unit is estimated with the help of the linear regression equation for each unit. In the second step, the classification of the units is performed by the mathematical programming model based on clustering analysis. The proposed method combines the strengths of regression analysis and mathematical programming method. From the 10 real data sets taken the well-known literature and simulation study results, it is observed that the proposed method outperforms the regression analysis, mathematical programming and artificial neural network based classification methods.