Analyse von Marktsegmenten mit Hilfe konnexionistischer Modelle

Hruschka, Harald and Natter, Martin (March 1992) Analyse von Marktsegmenten mit Hilfe konnexionistischer Modelle. Forschungsberichte / Research Memoranda 293

[img]
Preview
Text
fo293.pdf

Download (1MB) | Preview

Abstract or Table of Contents

Abstract: Success of market segmentation depends on the use of appropriate data analysis techniques. Connectionist models (artificial neural networks) constitute an alternative to well-known methods such as linear discriminant analysis or logistic regression. So-called backpropagation networks attain higher classification rates than logistic regression models in a pilot-study of a-priori segmentation. This may be caused by the capability of connectionist models for discovering nonlinear relationships between descriptors and segment memberships as well as interaction effects between descriptors. Special attention is drawn to specification of connectionist models and interpretation of results obtained from estimating their parameters.;

Item Type: IHS Series
Status: Published
Date Deposited: 26 Sep 2014 10:35
Last Modified: 01 Apr 2016 14:08
URI: http://irihs.ihs.ac.at/id/eprint/623

Actions (login required)

View Item View Item