Submitted by Sara Ribeiro (sara.ribeiro@ucb.br) on 2017-11-17T12:40:52Z No. of bitstreams: 1 AntonioBarbosaJuniorDissertacao2017.pdf: 3677259 bytes, checksum: 2e28ae94c001e36741ccb4fefefd5804 (MD5) / Approved for entry into archive by Sara Ribeiro (sara.ribeiro@ucb.br) on 2017-11-17T12:41:04Z (GMT) No. of bitstreams: 1 AntonioBarbosaJuniorDissertacao2017.pdf: 3677259 bytes, checksum: 2e28ae94c001e36741ccb4fefefd5804 (MD5) / Made available in DSpace on 2017-11-17T12:41:04Z (GMT). No. of bitstreams: 1 AntonioBarbosaJuniorDissertacao2017.pdf: 3677259 bytes, checksum: 2e28ae94c001e36741ccb4fefefd5804 (MD5) Previous issue date: 2017-08-31 / The evolution of Internet can be summarized in three phases. In the beginning, web 1.0 was marked by the World Wide Web, a network of static documents linked by the HTML markup language. Then it evolved into web 2.0, where bidirectional communi-cation enabled user interaction by means of social networking services, blogs, and wikis. Currently, the researches are focused on semantic web, or web 3.0, which has the purpose of marking the content of web in such a way that it is understandable by machines. Sensor network and wireless actuator technologies have also evolved to form a network of devices and objects connected to Internet that, in addition to sen-sors, also includes wearable technologies. For the theoretical basis of this work, a bibliographical research was carried out to access the most relevant publications on Internet of Things (IoT), discovery, repository and application discovery. The re-search showed that the Internet of Things paradigm presents some gaps. It was veri-fied that none of the researched approaches presents a proposal of repository to make a catalogue of applications, neither tools to discover applications and services of IoT for reuse. The purpose of this dissertation is to: (i) to propose a model of re-pository to catalogue IoT applications and services, and (ii) to develop a tool for dis-covering IoT applications and services for reuse. The goal is to turn the search and reuse of existing solutions easier in order to avoid rework and to speed up the devel-opment of applications for IoT. A set of relevant descriptions to the specification of the applications and the services, and a non-relational database for storing the infor-mation were used to create the repository model. To implement the discovery tool, Artificial Intelligence techniques were applied, specifically the kNN (Nearest Neigh-bor) classification algorithm, adapted to use the Levenshtein distance. The adapta-tion showed better results in identifying the nearest neighbor, that is, the application or service that complies better with the specifications and needs of the user. / A evolu????o da Internet pode ser sintetizada em tr??s fases. No in??cio, a web 1.0 foi marcada pela world wide web, uma rede de documentos est??ticos vinculados pela linguagem de marca????o HTML. Depois, ela evoluiu para a web 2.0, na qual a comu-nica????o bidirecional permitiu a intera????o do usu??rio com os servi??os de redes soci-ais, blogs e wikis. Atualmente, as pesquisas est??o voltadas para a web sem??ntica ou web 3.0, que tem a finalidade de marcar o conte??do da web de forma que seja com-preens??vel por m??quinas. As tecnologias de redes de sensores e atuadores sem fio tamb??m evolu??ram, e passaram a compor uma malha de dispositivos e objetos co-nectados na Internet que al??m dos sensores, abrange tamb??m as tecnologias vest??-veis Para a fundamenta????o te??rica deste trabalho, realizou-se uma pesquisa biblio-gr??fica com acesso ??s publica????es mais relevantes sobre Internet of Things, disco-very, repository e application discovery. Na pesquisa, constatou-se que o paradigma de Internet das Coisas apresenta algumas lacunas. Verificou-se que nenhuma das abordagens pesquisadas apresenta uma proposta de reposit??rio para catalogar apli-cativos, nem ferramentas para descoberta de aplicativos e servi??os de IoT para reu-so. A proposta desta disserta????o ??: (i) propor um modelo de reposit??rio para catalo-gar aplicativos e servi??os de IoT e, (ii) desenvolver uma ferramenta para descoberta de aplicativos e servi??os de IoT para reuso. O objetivo ?? facilitar a busca e o reuso de solu????es j?? existentes para evitar o retrabalho e agilizar o desenvolvimento de aplicativos para IoT. Para criar o modelo de reposit??rio, utilizou-se um conjunto de descritores relevantes para a especifica????o dos aplicativos e servi??os, e um banco de dados n??o relacional para o armazenamento das informa????es. Para implementar a ferramenta de descoberta, foram aplicadas t??cnicas de Intelig??ncia Artificial, mais especificamente o algoritmo de classifica????o kNN (Nearest Neighbor), adaptado pa-ra utilizar a dist??ncia de Levenshtein. A adapta????o apresentou melhores resultados para a identifica????o do vizinho mais pr??ximo, ou seja, aquele aplicativo ou servi??o mais aderente ??s especifica????es e necessidades do usu??rio.