Aluno: Paul Dany Flores Atauchi
Orientadora: Profª. Drª. Renata de Matos Galante
Coorientadora: Profª. Drª. Luciana Porcher Nedel
Título: Broker-RecSys: An Interactive Recommender System for Insurance Brokerage
Linha de pesquisa: Mineração, Integração e Análise de Dados
Esta banca ocorrerá excepcionalmente de forma totalmente remota. Interessados em assistir a defesa poderão acessar a sala virtual através do link: https://mconf.ufrgs.br/webconf/00109032
Prof. Dr. Silvio Cesar Cazella (UFCSPA – por videoconferência)
Profa. Dra. Carla Maria Dal Sasso Freitas (UFRGS – por videoconferência)
Prof. Dr. Leandro Krug Wives (UFRGS – por videoconferência)
Presidente da Banca: Profª. Drª. Renata de Matos Galante
Abstract: Recommender systems are widely used in more diverse domains such as e-commerce, tourism, insurance, and so on. However, numerous recommender systems do not take the user into account in the recommendation process and act like a black-box. Therefore, the black-box nature of the recommendation systems limits the understanding and acceptance of the recommendation received by the user. In contrast, interactive recommender systems can solve these drawbacks. Interactive recommender systems combine user interaction, information visualization, and recommender system methods. In the brokerage domain, insurance brokers offer, negotiate, and sell insurance products for their customers. Support brokers into the decision making for offering the most relevant insurance products for their customers can improve their loyalty, profit, and marketing campaign in their client portfolio. This work presents Broker-RecSys, an interactive insurance product recommender system to support brokers into the decision making for offering insurance products in their client portfolios. The system operates at two levels to provide recommendations: recommendations for a specific customer; and recommendations for a group of customers in the portfolio of clients of a broker. Looking for offering personalized recommendations, Broker-RecSys provides a module to perform customer segmentation based on specific customer characteristics that are interesting for the broker. Two types of recommendations are provided by Broker-RecSys: based on popularity and purchase behavior. Several interactions and visualization methods are integrated into Broker-RecSys in order to support brokers in the recommendation process. To evaluate Broker-RecSys, we combine the widely used evaluation method based on questionnaires and the evaluation based on the eye-tracking analysis. Broker-RecSys is evaluated into the usability and usefulness dimensions. Results achieved show that data mining methods, while combined with interaction and visualization methods, support users into the recommendation process and facilitate the decision-making to perform the recommendation of products.
Keywords: Interactive Recommender System. Data Mining. Decision Support System. Insurance Brokerage.