A Method for Credit Scoring Using Social Media

EXAMINERS:
Dr. Huang Ke-Wei, Department of Information Systems
Assoc Prof Stephane Bressan, Department of Computer Science

ABSTRACT:

While emerging economies have seen an explosion of social network site (SNS) adoption, these countries lack sophisticated credit bureaus to predict creditworthiness of individuals. In this paper, we propose an SNS-based credit scoring method for micro loans using large-scale observational data. We show empirical evidence that homophilous people interact with each other, and that after incorporating social network metrics, we can improve the repayment prediction by 30%. Next, we apply the semi-supervised classification for graph-based prediction. We show that SNS-based methods can be effective for developing countries that face the 'cold-start' problem.