Towards an Extension of Formal Concept Analysis Model for Peer-to-Peer Information Retrieval Systems

User profiles have significantly contributed to customizing user access and tailoring applications to individual needs. The challenge of automatically building user profiles remains a key research area, particularly in large-scale systems like Peer-to-Peer (P2P) networks. These systems are inherently distributed and lack a central structure to gather peer information, with each peer contributing independently. To enhance user satisfaction and application customization, user profiles can be leveraged. This paper proposes a user model that relies on historical search topics without external knowledge sources, learning from past interests to identify correlations between user requests, topics, documents, and nodes. The model extends formal concept analysis to cluster interactions into multidimensional concepts based on shared features such as topics and nodes. This approach aims to improve non-supervised tasks by incorporating usage data from past queries. We explore the integration of the proposed model in query routing and result aggregation for extensive distributed systems.