Author ORCID Identifier
Gelareh Towhidi: https://orcid.org/0009-0004-9088-3413
Atish P. Sinha: https://orcid.org/0000-0001-8125-7403
Huimin Zhao: https://orcid.org/0000-0002-6471-9837
Abstract
Prevailing computational models for link prediction in social networks are limited to "black boxes" that ignore socio-psychological trust antecedents and overlook the predictive value of distrust. To address this gap, we employ a Design Science Research (DSR) approach to develop a novel IT artifact: a theory-based predictive model for trust and distrust links. Grounded in kernel theories from social psychology and graph analysis, our model integrates features that capture trustor propensity, trustee prestige, and relational dynamics such as status differentials. Empirical evaluation demonstrates a 22% increase in accuracy and a 64% improvement in Matthews Correlation Coefficient over state-of-the-art baselines. The results confirm that theory-derived features provide superior predictive performance, thereby providing crucial explainability. Our contributions are threefold: (1) a design theory for explainable trust prediction; (2) a validated, high-performing artifact; and (3) practical insights for transparent trust-management systems. This work bridges computational and behavioral science, offering a holistic method for understanding online trust.
Recommended Citation
Towhidi, G., Sinha, A. P., & Zhao, H. (In press). A Theory-Based Link Prediction Model for Online Social Networks. Communications of the Association for Information Systems, 57, pp-pp. Retrieved from https://aisel.aisnet.org/cais/vol57/iss1/84
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.
