| Nome completo do aluno | Paulo Roberto de Souza |
|---|---|
| Título do trabalho | Exploiting Social Capital for Improving Personalized Recommendations in Online Social Networks |
| Resumo do trabalho | Online social networks generate vast volumes of content and user interactions, significantly intensifying information overload and limiting users’ ability to identify items that are both relevant and credible. This thesis addresses this challenge by proposing a recommender system that leverages social capital to enhance personalization in microblogging environments, using X (formerly Twitter) as the primary empirical setting. Social capital is operationalized as a measurable construct derived from observable network interactions and is computed from indicators of reputation, influence, and engagement strength. These signals are integrated with complementary attributes such as recency, diversity, and contextual alignment to determine ranking priority. The proposed approach constructs user profiles and content representations from public interaction traces, including likes, replies, retweets, mentions, and quotes, enabling the model to capture both topical interests and relational context. By jointly modeling content relevance and social signals, the system balances topical fit with network-aware cues of trust and influence, which are particularly salient in microblogging platforms where timeliness, credibility, and social endorsement are as important as semantic similarity. Grounding ranking decisions in network-derived attributes allows the recommender to better reflect how users assess usefulness, trustworthiness, and relevance in everyday social feeds. The effectiveness of the proposed model is evaluated through a comprehensive experimental framework that includes both an online user study and offline analyses conducted over the same dataset across three distinct experimental settings. The social-capital-aware algorithms are systematically compared against traditional and state-of-the-art baseline methods using standard ranking metrics, including MRR, MAP, NDCG, and Precision@N. Across all evaluation scenarios, the proposed methods consistently outperform the baselines, yielding measurable improvements in relevance and ranking quality while preserving model interpretability. Overall, the results demonstrate that incorporating social capital into the recommendation process provides a robust and practical means of mitigating information overload in online social networks. By delivering content that is not only relevant but also socially grounded and credible, the proposed approach establishes a reliable path toward the deployment of transparent and socially aware recommender systems in real-world. |
| Orientador | Frederico Araújo Durão |
| Membro Titular Externo 1 (com afiliação) | André Levi Zanon (Insight Research Ireland Centre for Data Analytics) |
| Link para o curriculum lattes |
http://lattes.cnpq.br/ |
| Membro Titular Externo 2 (com afiliação) | Natasha Correia Queiroz Lino (UFPB) |
| Link para o curriculum lattes |
http://lattes.cnpq.br/ |
| Membro Titular Interno 1 ou Titular Externo 3 (com afiliação) | Danilo Barbosa Coimbra (PGCOMP/UFBA) |
| Link para o curriculum lattes |
http://lattes.cnpq.br/ |
| Membro Titular Interno 2 ou Titular Externo 4 (com afiliação) | Laís do Nascimento Salvador (PGCOMP/UFBA) |
| Link para o curriculum lattes |
http://lattes.cnpq.br/ |
| Membro Suplente Externo 1 (com afiliação) | Renato de Freitas Bulcão Neto (UFG) |
| Link para o curriculum lattes |
http://lattes.cnpq.br/ |
| Membro Suplente Externo 2 (com afiliação) | Diogo Vinícius de Sousa Silva (IFMA) |
| Link para o curriculum lattes |
http://lattes.cnpq.br/ |
| Membro Suplente Interno 1 ou Suplente Externo 3 (com afiliação) | Rafael Augusto de Melo (PGCOMP/UFBA) |
| Link para o curriculum lattes |
http://lattes.cnpq.br/ |
| Membro Suplente Interno 2 ou Suplente Externo 4 (com afiliação) | Ivan do Carmo Machado (PGCOMP/UFBA) |
| Link para o curriculum lattes |
http://lattes.cnpq.br/ |
| Data da defesa | 04 Mar, 2026 |
| Horário da defesa | 8:30 AM |
| Quais os principais impactos deste trabalho (social, tecnológico, científico, ambiental)? | Our findings show tthat social capital plays a fundamental role in mediating social interactions, facilitating information flow, and fostering trust among individuals and groups. Its purposes are associated with enabling cooperation, strengthening social ties, providing access to informational resources, and enhancing the capacity of social actors to achieve individual and collective goals. In organizational and digital contexts, social capital contributes to reducing uncertainty, legitimizing content and decisions, and expanding the reach and credibility of shared information. In online social networks, these purposes are manifested through increased engagement, influence, and social recognition, serving as a mechanism that guides behavior, sustains long-term relationships, and enhances processes of recommendation, diffusion, and collaboration. |
Data da Defesa:
04/03/2026 - 08:30
Tipo de Defesa:
Defesa de Doutorado