Selection of multinomial logit models via association rules analysis

Pannapa Changpetch, Dennis K.J. Lin

Research output: Contribution to journalReview article

7 Citations (Scopus)

Abstract

In this research, we propose a novel approach for a multinomial logit model selection procedure: specifically, we apply association rules analysis to identifying potential interactions for multinomial logit modeling. Interaction effects are very common in reality, but conventional multinomial logit model selection methods typically ignore them. This is especially true for higher-order interactions. Here, we develop a model selection framework to address this problem. Specifically, we focus on building an optimal multinomial logit model by (1) exploring the combinations of input variables that have a significant impact on response (via association rules analysis); (2) selecting potential (low-order and high-order) interactions; (3) converting these potential interactions into new dummy variables; and (4) performing variable selections among all the input variables and the newly created dummy variables (interactions). Our model selection procedure is the first approach to provide a global search for potential interactions and establish the optimal combination of main effects and interaction effects in the multinomial logit model. In our investigation, we consider both simulated and real-life datasets, thereby confirming the effectiveness and efficiency of this method.

Original languageEnglish (US)
Pages (from-to)68-77
Number of pages10
JournalWiley Interdisciplinary Reviews: Computational Statistics
Volume5
Issue number1
DOIs
StatePublished - Jan 1 2013

Fingerprint

Multinomial Logit Model
Association Rules
Model Selection
Interaction
Interaction Effects
Selection Procedures
Multinomial Logit
Higher Order
Main Effect
Global Search
Variable Selection
Modeling

All Science Journal Classification (ASJC) codes

  • Statistics and Probability

Cite this

@article{ab7e05eb87704a7e8ec0ceff53147e68,
title = "Selection of multinomial logit models via association rules analysis",
abstract = "In this research, we propose a novel approach for a multinomial logit model selection procedure: specifically, we apply association rules analysis to identifying potential interactions for multinomial logit modeling. Interaction effects are very common in reality, but conventional multinomial logit model selection methods typically ignore them. This is especially true for higher-order interactions. Here, we develop a model selection framework to address this problem. Specifically, we focus on building an optimal multinomial logit model by (1) exploring the combinations of input variables that have a significant impact on response (via association rules analysis); (2) selecting potential (low-order and high-order) interactions; (3) converting these potential interactions into new dummy variables; and (4) performing variable selections among all the input variables and the newly created dummy variables (interactions). Our model selection procedure is the first approach to provide a global search for potential interactions and establish the optimal combination of main effects and interaction effects in the multinomial logit model. In our investigation, we consider both simulated and real-life datasets, thereby confirming the effectiveness and efficiency of this method.",
author = "Pannapa Changpetch and Lin, {Dennis K.J.}",
year = "2013",
month = "1",
day = "1",
doi = "10.1002/wics.1242",
language = "English (US)",
volume = "5",
pages = "68--77",
journal = "Wiley Interdisciplinary Reviews: Computational Statistics",
issn = "1939-5108",
publisher = "John Wiley and Sons Inc.",
number = "1",

}

Selection of multinomial logit models via association rules analysis. / Changpetch, Pannapa; Lin, Dennis K.J.

In: Wiley Interdisciplinary Reviews: Computational Statistics, Vol. 5, No. 1, 01.01.2013, p. 68-77.

Research output: Contribution to journalReview article

TY - JOUR

T1 - Selection of multinomial logit models via association rules analysis

AU - Changpetch, Pannapa

AU - Lin, Dennis K.J.

PY - 2013/1/1

Y1 - 2013/1/1

N2 - In this research, we propose a novel approach for a multinomial logit model selection procedure: specifically, we apply association rules analysis to identifying potential interactions for multinomial logit modeling. Interaction effects are very common in reality, but conventional multinomial logit model selection methods typically ignore them. This is especially true for higher-order interactions. Here, we develop a model selection framework to address this problem. Specifically, we focus on building an optimal multinomial logit model by (1) exploring the combinations of input variables that have a significant impact on response (via association rules analysis); (2) selecting potential (low-order and high-order) interactions; (3) converting these potential interactions into new dummy variables; and (4) performing variable selections among all the input variables and the newly created dummy variables (interactions). Our model selection procedure is the first approach to provide a global search for potential interactions and establish the optimal combination of main effects and interaction effects in the multinomial logit model. In our investigation, we consider both simulated and real-life datasets, thereby confirming the effectiveness and efficiency of this method.

AB - In this research, we propose a novel approach for a multinomial logit model selection procedure: specifically, we apply association rules analysis to identifying potential interactions for multinomial logit modeling. Interaction effects are very common in reality, but conventional multinomial logit model selection methods typically ignore them. This is especially true for higher-order interactions. Here, we develop a model selection framework to address this problem. Specifically, we focus on building an optimal multinomial logit model by (1) exploring the combinations of input variables that have a significant impact on response (via association rules analysis); (2) selecting potential (low-order and high-order) interactions; (3) converting these potential interactions into new dummy variables; and (4) performing variable selections among all the input variables and the newly created dummy variables (interactions). Our model selection procedure is the first approach to provide a global search for potential interactions and establish the optimal combination of main effects and interaction effects in the multinomial logit model. In our investigation, we consider both simulated and real-life datasets, thereby confirming the effectiveness and efficiency of this method.

UR - http://www.scopus.com/inward/record.url?scp=84871831057&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84871831057&partnerID=8YFLogxK

U2 - 10.1002/wics.1242

DO - 10.1002/wics.1242

M3 - Review article

AN - SCOPUS:84871831057

VL - 5

SP - 68

EP - 77

JO - Wiley Interdisciplinary Reviews: Computational Statistics

JF - Wiley Interdisciplinary Reviews: Computational Statistics

SN - 1939-5108

IS - 1

ER -