Desenvolvimento de um sistema para gerenciamento de catálogos pessoais de jogos digitais

Abstract

The growth of the digital gaming market and the widespread availability of titles make organizing personal libraries and discovering new games increasingly challenging tasks for users. In this context, this work presents the development of a web system for managing personal catalogs of digital games, capable of integrating library organization functionalities with a recommendation engine based on large-scale language models. The solution was developed using Java, Spring Boot, Hibernate, Flyway and Post- greSQL, in addition to integration with the RAWG API to obtain game information and with the Groq platform to generate recommendations. Initially, the system requirements were raised and the models that guided its implementation were developed. Next, a REST API was developed with features for user authentication, profile management, game library search and organization, and the generation of personalized recommen- dations from a title or natural language description. The application was validated through functional testing of the API endpoints and recommendation quality assessment using the LLM as a Judge approach based on criteria related to similarity of experi- ence, gameplay, setting, justifications and overall quality of recommendations. The results obtained demonstrate that the developed solution met the proposed objectives, presenting satisfactory performance both in implementing the planned functionalities and in generating recommendations consistent with the preferences informed by users. Thus, the work highlights the feasibility of using language models to support game recommendation systems, contributing to the organization of digital libraries and the discovery of new titles.