A Branch-and-Price approach to find optimal decision trees

M. Firat, Guillaume Crognier, Adriana Gabor, Y. Zhang

Research output: Contribution to conferenceAbstractAcademic

Abstract

In Artificial Intelligence (AI) field, decision trees have gained certain importance due to their effectiveness in solving classification and regression problems. Recently, in the literature we see finding optimal decision trees are formulated as Mixed Integer Linear Programming (MILP) models. This elegant way enables us to construct optimal decision trees with different error concerns. In our work, we find the optimal decision trees by employing the Column Generation (CG) approach. To do so, we first reformulated the previously proposed MILP model as a master MILP model. This master model is basically a partitioning problem of the rows in the given data set such that every We obtain the integer solutions by applying a Branch-and-Price search. Our approach of constructing decision trees is successfully tested in real-world benchmark datasets.
Original languageEnglish
Publication statusPublished - 8 Jul 2018

Fingerprint

Decision trees
Linear programming
Artificial intelligence

Keywords

  • Combinatorial Optimization
  • Artificial Intelligence
  • Analytics and Data Science

Cite this

@conference{1ed2c32451d34658a2b4a4f4afc31a79,
title = "A Branch-and-Price approach to find optimal decision trees",
abstract = "In Artificial Intelligence (AI) field, decision trees have gained certain importance due to their effectiveness in solving classification and regression problems. Recently, in the literature we see finding optimal decision trees are formulated as Mixed Integer Linear Programming (MILP) models. This elegant way enables us to construct optimal decision trees with different error concerns. In our work, we find the optimal decision trees by employing the Column Generation (CG) approach. To do so, we first reformulated the previously proposed MILP model as a master MILP model. This master model is basically a partitioning problem of the rows in the given data set such that every We obtain the integer solutions by applying a Branch-and-Price search. Our approach of constructing decision trees is successfully tested in real-world benchmark datasets.",
keywords = "Combinatorial Optimization, Artificial Intelligence, Analytics and Data Science",
author = "M. Firat and Guillaume Crognier and Adriana Gabor and Y. Zhang",
year = "2018",
month = "7",
day = "8",
language = "English",

}

A Branch-and-Price approach to find optimal decision trees. / Firat, M.; Crognier, Guillaume; Gabor, Adriana; Zhang, Y.

2018.

Research output: Contribution to conferenceAbstractAcademic

TY - CONF

T1 - A Branch-and-Price approach to find optimal decision trees

AU - Firat, M.

AU - Crognier, Guillaume

AU - Gabor, Adriana

AU - Zhang, Y.

PY - 2018/7/8

Y1 - 2018/7/8

N2 - In Artificial Intelligence (AI) field, decision trees have gained certain importance due to their effectiveness in solving classification and regression problems. Recently, in the literature we see finding optimal decision trees are formulated as Mixed Integer Linear Programming (MILP) models. This elegant way enables us to construct optimal decision trees with different error concerns. In our work, we find the optimal decision trees by employing the Column Generation (CG) approach. To do so, we first reformulated the previously proposed MILP model as a master MILP model. This master model is basically a partitioning problem of the rows in the given data set such that every We obtain the integer solutions by applying a Branch-and-Price search. Our approach of constructing decision trees is successfully tested in real-world benchmark datasets.

AB - In Artificial Intelligence (AI) field, decision trees have gained certain importance due to their effectiveness in solving classification and regression problems. Recently, in the literature we see finding optimal decision trees are formulated as Mixed Integer Linear Programming (MILP) models. This elegant way enables us to construct optimal decision trees with different error concerns. In our work, we find the optimal decision trees by employing the Column Generation (CG) approach. To do so, we first reformulated the previously proposed MILP model as a master MILP model. This master model is basically a partitioning problem of the rows in the given data set such that every We obtain the integer solutions by applying a Branch-and-Price search. Our approach of constructing decision trees is successfully tested in real-world benchmark datasets.

KW - Combinatorial Optimization

KW - Artificial Intelligence

KW - Analytics and Data Science

M3 - Abstract

ER -