Prototipo de aplicación móvil para la recomendación automática de una orientación psicoterapéutica a través del uso de Machine Learning

Abstract

The main purpose of this project is to develop a prototype mobile application aimed at recommending the most appropriate psychotherapeutic orientation for the individual needs of users using natural language processing (NLP) and Machine Learning techniques with Transformers and BERT. The challenge lies in the users' lack of knowledge and understanding of the various psychotherapeutic orientations, which can lead to inappropriate choices for their condition. By using the SCRUM methodology involved the division of processes into specific activities for the development of the modules that are implemented in the mobile application prototype, highlighting the chatbot as the main element, during the interaction with the user by using the OpenAI service and subsequently saving the conversation in a non-relational database. The implemented architecture is supported by an n-layer structure that includes a data layer with a non-relational database management platform, a business layer with Express and Flask servers to send and receive the data, and the user interface or presentation layer developed in React-Native. The results obtained showed that the prototype, especially the chatbot module, had a positive acceptance by the users. In conclusion, this work contributes to the solution of the identified problem, providing users with a tool that facilitates the appropriate choice of psychotherapeutic guidance, based on the analysis of a conversion between the user and the chatbot.