Linguistic Features and Loan Quality: A Machine Learning Approach

Our research studies the role of texts in Crowdfunding. Borrowers write texts when they request funds on, one leading lending-based Crowdfunding site. We study whether linguistic features of these texts, or how they are written, can indicate the creditworthiness of the borrowers. For that purpose, we use both explanatory and predictive modeling approaches. We also take one step further to see how investors actually responded to linguistic features.

Session 5A: Crowd Computing, Paper# 1