C-DeepTrust: A Context-Aware Deep Trust Prediction Model in Online Social Networks

Area of Science:

• Computer Science
• Social Network Analysis
• Machine Learning

Background:

• Trust prediction is crucial for online social networks, influencing decisions, information spread, and marketing.
• Existing models often overlook key factors like interaction behavior, context, and temporal dynamics in trust establishment.
• A comprehensive approach is needed to model user trust development holistically.

Purpose of the Study:

• To propose C-DeepTrust, a context-aware deep trust prediction model.
• To integrate interaction behaviors, contextual information, and temporal features for enhanced trust prediction.
• To address the limitations of existing trust prediction methods in online social networks.

Main Methods:

• User feature modeling to extract static and dynamic preferences within specific contexts.
• Utilizing a multilayer projection structure to mine context-aware latent features.
• Calculating trust relationships via cosine similarity of context-aware feature vectors, grounded in social homophily theory.

Main Results:

• C-DeepTrust demonstrates superior performance compared to existing baseline methods.
• The model effectively incorporates diverse data types for more accurate trust prediction.
• Experiments on real-world datasets validate the model's effectiveness.

Conclusions:

• The proposed C-DeepTrust model offers a significant advancement in trust prediction for online social networks.
• Integrating context and temporal dynamics leads to more robust and accurate trust relationship modeling.
• The findings support the application of C-DeepTrust in various online social network scenarios.