| Name: | Description: | Size: | Format: | |
|---|---|---|---|---|
| 9.04 MB | Adobe PDF |
Authors
Advisor(s)
Abstract(s)
Os sistemas de recomendação têm valorizado e aumentado em popularidade nos últimos anos, com os utilizadores a procurarem a facilidade e simplicidade na obtenção de informações e sugestões, tendo por base as suas preferências e expectativas. Adicionalmente, o estilo de vida fitness está na moda com cada vez mais pessoas a adotá-lo. Dessa forma, os praticantes deste tipo de estilo de vida utilizam aplicações para registar as suas atividades, definir objetivos, monitorizar o progresso e evolução, entre outros. Contudo, a maioria das aplicações são específicas a um tipo de utilizador, sendo difícil encontrar uma que seja direcionada a praticantes iniciantes e, ao mesmo tempo, a experientes. Além disso, existem várias lacunas nas aplicações existentes, como falta de funcionalidades importantes ou aspetos de usabilidade. Posto isto, foi desenvolvido um sistema com uma componente móvel que permite a qualquer tipo de praticante de fitness realizar as funcionalidades fundamentais que a sua atividade requer, tais como registo das suas atividades, consulta de treinos ou exercícios fitness já existentes ou criados, personalização de planos e esclarecimento de dúvidas com um assistente pessoal. Além disso, possui um módulo de recomendação com o objetivo de recomendar treinos com base na natureza destes (p.e. nível ou músculos exercitados), nos objetivos do mesmo, e nas características do utilizador. Para validar as funcionalidades gerais, a utilidade e a usabilidade dos componentes desenvolvidos, foram elaborados inquéritos de satisfação. O resultado foi bastante positivo, onde uma elevada percentagem dos utilizadores inquiridos indicou que os módulos desenvolvidos eram bastante úteis, intuitivos e simples. Adicionalmente, as técnicas de recomendação também foram alvo de teste de forma a analisar o erro de previsão de cada uma delas. De forma geral, os resultados também foram positivos, com erros de previsão consideravelmente baixos.
Recommender Systems have been increasing in popularity and appreciation over the past few years, with the users seeking ease and simplicity when obtaining information and suggestions, based on their preferences and expectations. Additionally, the fitness lifestyle is being adopted by increasingly more people. In this way, practitioners of this type of lifestyle use applications to record their activities, set goals, monitor progress and evolution, among others. However, most applications are specific to one type of user, and it is difficult to find one that is targeted at both beginner and experienced practitioners. Moreover, there are several gaps in existing applications, such as lack of important functionalities and usability aspects. With this, a mobile component system has been developed, which allows any type of fitness practitioner to perform the fundamental features that their activity requires, such as recording them, consulting existing or created fitness workouts, customizing plans and questions with a personal assistant. In addition, it has a recommendation module with the purpose of recommending workouts based on their nature (i.e. experience level or muscles exercised), and the user's characteristics and objectives. To validate the general functionality, usefulness and usability of the developed components, satisfaction inquiries were developed. The result was very positive, where a high percentage of users stated that the modules developed were very useful, intuitive and simple. Additionally, the recommendation techniques were also tested to analyze the prediction error of each one. Overall, the results were also positive, with considerably low errors.
Recommender Systems have been increasing in popularity and appreciation over the past few years, with the users seeking ease and simplicity when obtaining information and suggestions, based on their preferences and expectations. Additionally, the fitness lifestyle is being adopted by increasingly more people. In this way, practitioners of this type of lifestyle use applications to record their activities, set goals, monitor progress and evolution, among others. However, most applications are specific to one type of user, and it is difficult to find one that is targeted at both beginner and experienced practitioners. Moreover, there are several gaps in existing applications, such as lack of important functionalities and usability aspects. With this, a mobile component system has been developed, which allows any type of fitness practitioner to perform the fundamental features that their activity requires, such as recording them, consulting existing or created fitness workouts, customizing plans and questions with a personal assistant. In addition, it has a recommendation module with the purpose of recommending workouts based on their nature (i.e. experience level or muscles exercised), and the user's characteristics and objectives. To validate the general functionality, usefulness and usability of the developed components, satisfaction inquiries were developed. The result was very positive, where a high percentage of users stated that the modules developed were very useful, intuitive and simple. Additionally, the recommendation techniques were also tested to analyze the prediction error of each one. Overall, the results were also positive, with considerably low errors.
Description
Keywords
Fitness Exercícios Monitorização Sistemas de Recomendação Assistente Pessoal Exercises Monitoring Recommender Systems Personal Virtual Assistant