Sistema de recomendação para alívio de stress

• Autor(a) principal: Pinelas, Francisco Miguel Magalhães
• Data de Publicação: 2022
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10400.22/21485
• topic: Stress; Recommender system; Recommendations; Filtering techniques; Contentbased filtering; Sistema de recomendação; Recomendações; Técnicas de filtragem; Filtragem baseada em conteúdo

Stress is considered to be a normal part of our lives, especially when taken into account that people are constantly trying to push their limits and the limits of others around them. Whether at home or at their jobs, the idea that to be successful one must work harder is deeply rooted within society because such behavior has shown positive results in the past. Stressful events can work as a reactor for people to feel the necessary motivation to move on with their tasks. However, if uncontrolled, may lead to health-related consequences, such as cardiovascular diseases, sleep deprivation, and anxiety. Therefore, it is important to not only recognize that stress has serious negative impacts in the lives of people but also to find mechanisms to cope with it. This document presents a solution in the form of a web application capable of providing stresseasing and health improving recommendations that adapt to the users of the application by considering their ratings from past recommendation, as well as their profiles. Through an engaging and interactive graphical user interface, users can receive personalized recommendations via system notifications. These notifications are composed of a demonstrative card and a description, as well as a collection of documents with insightful information regarding the recommendations being provided. This web application is supported by three major components, designed to operate both synchronously and asynchronously in a microservices-oriented architecture, promoting the flexibility and scalability of the solution. Furthermore, for the solution to provide recommendations, it was necessary to implement a filtering technique. Among the most common ones, the content-based filtering is the most advantageous, meaning that a content-based recommender system was developed as part of the solution. Lastly, it was concluded that the web application satisfactorily meets the established requirements. However, due to lack of user-generated data, the randomly generated data used to demonstrate how a proper evaluation would be conducted cannot be subject for interpretation.