| Título: | A framework for prediction of non-aeronautical revenues based on customer experience using computational intelligence hybrid models |
| Autor: | Lenice Mirian da Silva |
| Programa: | Engenharia de Infraestrutura Aeronáutica |
| Área de Concentração: | Transporte Aéreo e Aeroportos |
| Orientador : | Giovanna Miceli Ronzani Borille |
| Ano de Publicação : | 2022 |
| Curso : | Doutorado |
| Assuntos : | Aeroportos |
| t | Centros comerciais |
| t | Qualidade de serviço |
| t | Análise envoltória de dados |
| t | Redes neurais |
| t | Transportes |
| Resumo : | In an airport system, revenues are divided into aeronautical and non-aeronautical, where one of the aeronautical revenues is linked to boarding fees while the non-aeronautical revenues are related to commercial services. Commercial activities impact airport development since commercial indicators make up the Service Quality Indicators and are associated with the Passenger Satisfaction Survey. In this sense, the customer experience in commercial establishments is part of the passenger journey at the airport. This research proposes a study on the full exploitation of the commercial area of the airport considering two segments of non-aeronautical revenue: retail and food and beverage. The academic contributions of this research in relation to existing studies are: (i) an analysis of the commercial performance of the busiest airport in Brazil and a brief comparison with other airports in the world, (ii) an analysis of the mix of stores, (iii) a neural network proposal for passenger satisfaction prediction based on passenger profile, and (iv) a model for predicting the amount spent per passenger. The framework developed is a combination of intelligent computational models that include Deep Artificial Neural Network and Fuzzy Logic techniques. Multilayer Neural Networks are used for prediction problems where it learns from large historical data and generalizes to predicted data, while Fuzzy Logic is used to address problems involving human behavior. The results found in this research show that it is possible to investigate passenger consumption through a hybrid computational model structured by a proposed framework that considers non-aeronautical revenues and customer experience. |
| Data de Defesa : | 13/12/2022 |
| Texto na íntegra : | [Visualizar] |