Download full text
(2.029Mb)
Citation Suggestion
Please use the following Persistent Identifier (PID) to cite this document:
https://nbn-resolving.org/urn:nbn:de:0168-ssoar-49482-6
Exports for your reference manager
Nonparametric modeling of buying behavior in fast moving consumer goods markets
Nichtparametrische Modellierung des Kaufverhaltens auf Märkten für kurzlebige Konsumgüter
[conference paper]
Corporate Editor
Zentrum für Umfragen, Methoden und Analysen -ZUMA-
Abstract "From the statistical point of view a nonparametric formulation of a brand choice model (NDE) is a powerful alternative to the logit model. But in the marketing context, researchers in general want to have parameter values to make predictions or to estimate market shares. This leads to a semiparamet... view more
"From the statistical point of view a nonparametric formulation of a brand choice model (NDE) is a powerful alternative to the logit model. But in the marketing context, researchers in general want to have parameter values to make predictions or to estimate market shares. This leads to a semiparametric model (GAM) formulation with two possible ways of using the results. One is to perform estimation of choice probabilities, but there one is confronted with the same problem as in the nonparametric approach, because no parameters are estimated for the nonparametric part of the model. The second possibility of a semiparametric model formulation overcomes this problem. In addition, with the estimation results a modified parametric model formulation can be estimated. This also gives the possibility to work with the parameter values to estimate market shares or make predictions. Especially for this use of modeling, the underlying data structure should be detected correctly. Therefore, two different estimation algorithms for a GAM were presented and the application of the semiparametric model to a real data set was reported. The estimations were made by the two common algorithms, backfitting and marginal integration, and are compared to each other. An interaction effect in the variable price in the data set was discovered, which leads to the need of additional studies of the data set." (author's abstract)... view less
Keywords
consumer; consumer goods; buying behavior; economic model; data; statistical analysis; research approach; method; estimation; non-linear model; parameter; market; Federal Republic of Germany
Classification
Methods and Techniques of Data Collection and Data Analysis, Statistical Methods, Computer Methods
National Economy
Method
empirical; quantitative empirical
Collection Title
Social and economic research with consumer panel data : proceedings of the first ZUMA Symposium on Consumer Panel Data, 5 and 6 October 1999
Editor
Papastefanou, Georgios; Schmidt, Peter; Börsch-Supan, Axel; Lüdtke, Hartmut; Oltersdorf, Ulrich
Conference
1. ZUMA Symposium on Consumer Panel Data. Mannheim, 1999
Document language
English
Publication Year
2001
City
Mannheim
Page/Pages
p. 189-205
Series
ZUMA-Nachrichten Spezial, 7
ISBN
3-924220-21-2
Status
Published Version; reviewed
Licence
Deposit Licence - No Redistribution, No Modifications