Modelos aditivos generalizados para a avaliação da intenção de compra de consumidores

• Autor(a) principal: Souza, Erivaldo Lopes de
• Data de Publicação: 2012
• Tipo de documento: Dissertação
• Idioma: por
• Título da fonte: Biblioteca Digital de Teses e Dissertações da UFPB
• Texto Completo: https://repositorio.ufpb.br/jspui/handle/tede/5232
• topic: Comportamento do consumidor; Intenção de compra; Modelos Aditivos Generalizados; Segmentação de mercado; Consumer behavior; Purchase intent; Generalized Additive Models; Market segmentation; CNPQ::ENGENHARIAS::ENGENHARIA DE PRODUCAO

In recent years, several studies have been published dealing on factors that influence consumer purchase intent in various economic sectors. In this line of work, we are specifically for the sector of collective buying, obtain regression models that could contribute to the study of the relationship between the purchase intention and the characteristics of market segments. The aim is to assist in the inclusion of this variable purchase intention in the process of choosing a target group, guiding decisions to meet more effectively service consumers. To achieve this goal, initially interviewed 384 Internet users in the city of João Pessoa, Paraíba, Brazil. Then the data obtained from interviews, were used to estimate those models. These models were based on assumptions of theories of the cognitive approach of consumer behavior, especially the Theory of Reasoned Action. The instrument used for data collection was a questionnaire containing market research questions related to psychological factors, socio-cultural and situational consumer. The most successful model was a generalized additive model with nine variables and nonparametric one end, obtained from smoothing splines. This model had a pseudo-R2 of 0,89 and allowed to reach a percentage of correct trials of the observations of the sample equal to 94%. With the aid of simulations, it was observed how the proposed model type is capable of assisting in selection of a target with a higher interest in the use of the service. It was also shown how the model can be used to evaluate production systems, in relation to more efficient service to customers intend to use the service. The generalized additive models were effective for identifying the presence of nonlinear relationships and were able to generate a high explanatory power of the propensity of individuals to use specific service.