Proceedings of The 5th International Conference on Modern Research in Engineering, Technology and Science
Digital Service in Brazilian Large Collections: Proposed Mobile Solution Integrated to Chatbots applied to University Collections
Alan Erick Eremita Dias, Jéssica Paula Costa Duarte, Leandro José de Oliveira Gama, José Humberto Cruvinel Resende Junior, Vladimir Alexei Rodrigues Rocha, Gustavo Alves Fernandes, Bráulio Roberto Gomes Marinho Couto, Diva de Souza e Silva Rodrigues, Luiz Melk de Carvalho and Flávio Henrique Batista de Souza
This article describes the initial process of developing a system for managing personalized service processes in the physical spaces of the library in higher education networks in Brazil, with its operation focused on accredited networks in the rental of books to the community. The development of the mobile customer service application, called chatbot, will assist on the evaluation process of libraries, based on artificial intelligence, with deep learning techniques and recurrent neural networks for the definition of the profile of the user. With the proposed tool an advanced analysis of emotions is performed to provide recommendations to the user about their tasks, according to the level of didactic instructions in the book. The implementations of the structures and the experiments to train the tool to learn the demands of the user were carried out. As contributions of this work it can be listed: the demonstration of the database treatment process with real information from the queries of a library for NLP (Natural Language Processing); the implementation of the web structure with the support of a chatbot to provide guidance for the user; the demonstration of experiments (and challenges faced) with tools based on Artificial Neural Networks in their use for learning about the profile of the user during the conversation with the chatbot; and an analysis of future feasibility of integration in literary services.
keywords: University library, Chatbot, Virtual assistant, Natural language processing, Artificial Neural Networks.