APPLICATION OF DATA MINING ALGORITHM FOR REPRESENTING NON-COMPENSATORY DECISION STRATEGY

DOI IR HANDLE Web Site Web Site 1 Citations 59 References Open Access
  • YAMAMOTO Toshiyuki
    社団法人 土木学会 名古屋大学大学院 工学研究科社会基盤工学専攻

Bibliographic Information

Other Title
  • 非補償型意思決定方略を表現するためのデータマイニング手法の適用に関する分析
  • ヒホショウガタ イシ ケッテイ ホウリャク オ ヒョウゲン スル タメ ノ データマイニング シュホウ ノ テキヨウ ニ カンスル ブンセキ

Search this article

Abstract

C4.5, one of the popular data mining algorithms, is applied to represent non-compensatory choice behavior, and the predictability of the choice behavior is compared with the conventional logit model. The empirical analysis of this study is based on the SP data about the use of a hypothetical dynamic park and ride system. The results suggest that C4.5 and the conventional model have a same predictability in terms of the hit ratio. Though, the log-likelihood at convergence of the logit model with the dummy variables representing if-then rules produced by C4.5 is lower than that of the conventional logit model. When the dummy variables are added to the conventional logit model with conventional independent variables, the higher log-likelihood at convergence than that of the conventional logit model is given even after the insignificant variables are excluded.

Journal

Citations (1)*help

See more

References(59)*help

See more

Related Projects

See more

Keywords

Details 詳細情報について

Report a problem

Back to top