Loading...
Publication
Sistema de RecomendaĆ§Ć£o Nutricional para Pacientes com Diabetes Mellitus Tipo 2
| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 2.8 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
A diabetes mellitus tipo 2 Ć© considerada uma das doenƧas crĆ³nicas mais comuns no mundo. Atualmente ainda nĆ£o existe uma cura, mas o tratamento assenta num controlo e acompanhamento contĆnuo da mesma. A autogestĆ£o da doenƧa faz parte de uma das soluƧƵes previstas, uma vez que pode incentivar Ć adoĆ§Ć£o de hĆ”bitos saudĆ”veis. SĆ£o vĆ”rios os mĆ©todos que podem ser praticados nesta autogestĆ£o, nomeadamente mĆ©todos tradicionais que estĆ£o dependentes do uso de papel e caneta, como diĆ”rios de alimentos, que se revelam pouco eficazes e nĆ£o correspondem, sempre, Ć s necessidades do utilizador. Isto leva Ć necessidade de desenvolver sistemas que mitiguem esses problemas e que sejam capazes de auxiliar o paciente no controlo da doenƧa no seu dia a dia. Sendo assim, faz sentido estudar uma abordagem que permita ao utilizador registar o seu perfil, a sua rotina diĆ”ria obtendo, por fim, recomendaƧƵes tendo em conta esses dados. Para atingir esse objetivo foi desenvolvido, no trabalho descrito nesta dissertaĆ§Ć£o, um sistema de recomendaĆ§Ć£o, associado a uma aplicaĆ§Ć£o mĆ³vel, que visa proporcionar ao paciente uma autogestĆ£o da doenƧa em questĆ£o. Este sistema, recomenda ao utilizador alimentos, de acordo com o seu registo diĆ”rio e com as suas preferĆŖncias alimentares, adaptando-se Ć s necessidades do mesmo. Para o desenvolvimento do sistema de recomendaĆ§Ć£o, optou-se por um algoritmo cuja tĆ©cnica abordada Ć© baseada em filtragem colaborativa, sendo que considera os alimentos que o utilizador classificou, mas tambĆ©m classificaƧƵes atribuĆdas por outros utilizadores, apresentando-os como uma possĆvel soluĆ§Ć£o. Para perceber os comportamentos do sistema desenvolvido, foram definidos casos de teste para analisar os diferentes cenĆ”rios de um dia a dia do paciente, no que diz respeito aos alimentos que ingere, e aos valores nutritivos dos mesmos. Por fim, os resultados obtidos permitem concluir que o sistema desenvolvido recomenda de acordo com as preferĆŖncias do utilizador, mas tambĆ©m de outros atravĆ©s das classificaƧƵes que foram atribuĆdas aos alimentos.
Self-management of the disease is part of one of the solutions provided since it can encourage the adoption of healthy habits. There are several methods that can be practiced in this self-management, namely traditional methods that are dependent on the use of paper and pen, such as food diaries, which prove to be ineffective and do not always correspond to the user's needs. This leads to the need to develop systems that mitigate these problems and that can assist the patient in controlling the disease daily. Therefore, it makes sense to study an approach that allows the user to register their profile, their daily routine, finally obtaining recommendations considering these data. To achieve this goal, a recommender system was developed in the work described in this dissertation, associated with a mobile application, which aims to provide the patient with self-management of the disease in question. This system recommends food to the user, according to his daily record and his food preferences, adapting to his needs. For the development of the recommender system, we opted for an algorithm whose technique is based on collaborative filtering, considering the foods that the user has classified, but also ratings assigned by other users, presenting them as a possible solution. In order to understand the behaviour of the developed system, test cases were defined to analyse the different scenarios of the patient's daily life, with regard to the food he eats, and their nutritional values. Finally, the results obtained allow us to conclude that the developed system recommends according to the user's preferences, but also from others, through the classifications that have been attributed to food. Type 2 diabetes mellitus is considered one of the most common chronic diseases in the world. There is currently no cure, but the treatment is based on continuous control and monitoring.
Self-management of the disease is part of one of the solutions provided since it can encourage the adoption of healthy habits. There are several methods that can be practiced in this self-management, namely traditional methods that are dependent on the use of paper and pen, such as food diaries, which prove to be ineffective and do not always correspond to the user's needs. This leads to the need to develop systems that mitigate these problems and that can assist the patient in controlling the disease daily. Therefore, it makes sense to study an approach that allows the user to register their profile, their daily routine, finally obtaining recommendations considering these data. To achieve this goal, a recommender system was developed in the work described in this dissertation, associated with a mobile application, which aims to provide the patient with self-management of the disease in question. This system recommends food to the user, according to his daily record and his food preferences, adapting to his needs. For the development of the recommender system, we opted for an algorithm whose technique is based on collaborative filtering, considering the foods that the user has classified, but also ratings assigned by other users, presenting them as a possible solution. In order to understand the behaviour of the developed system, test cases were defined to analyse the different scenarios of the patient's daily life, with regard to the food he eats, and their nutritional values. Finally, the results obtained allow us to conclude that the developed system recommends according to the user's preferences, but also from others, through the classifications that have been attributed to food. Type 2 diabetes mellitus is considered one of the most common chronic diseases in the world. There is currently no cure, but the treatment is based on continuous control and monitoring.
Description
Keywords
Sistemas de RecomendaĆ§Ć£o Filtragem Colaborativa Diabetes Mellitus Tipo 2 Recommender Systems Collaborative Filtering Type 2 Diabetes Mellitus