A quadratic fuzzy regression approach for handling uncertainties in Partial Least Squares Path Modeling

In marketing and education research it is usual to use Structural Equation Modeling (SEM) methods to fit network of constructs to data collected by surveys. Lately an SEM approach known as Partial Least Squares Path Modeling that tries to maximize the total explained variance of the variables has be...

Full description

Saved in:
Bibliographic Details
Published in:Revista IEEE América Latina 2018-01, Vol.16 (1), p.192-201
Main Authors: Seman, Laio Oriel, Gomes, Giancarlo, Hausmann, Romeu, Bezerra, Eduardo Augusto
Format: Article
Language:English
Subjects:
Computational modeling
Data models
Fuzzy Regression
IEEE transactions
Mathematical model
Numerical analysis
Partial Least Squares Path Modeling
Possibilistic regression
Quadratic Optimization
RNA
Structural Equation Modeling
Uncertainty
Citations: Items that cite this one
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:In marketing and education research it is usual to use Structural Equation Modeling (SEM) methods to fit network of constructs to data collected by surveys. Lately an SEM approach known as Partial Least Squares Path Modeling that tries to maximize the total explained variance of the variables has been widely applied. In those areas, the collected data may contain uncertainty of the human condition, which can bias the analyzed results. Lately, a Partial Least Squares Path Modeling algorithm with fuzzy regression was proposed in the literature to handle those uncertainties, using a possibilistic regression that has been suffering critics. This paper proposes a modified approach for handling data uncertainty in Partial Least Squares Path Modeling by means of a quadratic fuzzy regression with central tendency. The proposed approach is tested in the well know ECSI model to compare the obtained results with the traditional literature.
ISSN:1548-0992
1548-0992
DOI:10.1109/TLA.2018.8291473