Abstract: P2P lending let borrowers to obtain trust of investors through a description-text of borrower. So the description-text of borrower is an important information for investors to identify default risk of borrowers. However, how to interpret description-text with the complex, irregular and contain various kinds of information faces great challenges. According to the two factors of default risk:repayment ability and willingness to repay, as well as their potential factors, some text features,information of repayment ability and willingness to repay as well as emotional characteristics of the demand for funds are extracted from description-texts through manual identification, and the significance of those information to identify the default risk of borrowers is tested. It is found that the more words in the description-text, the more repetition sentences existing, and the default risk is greater. There is information of repayment ability or the guarantee language indicating the willingness to repay the loan and the supplementary explanation of the credit status in the description-text, the default risk is smaller. The greater the urgency of the borrower's emotionally expressed need for the loan, the greater the default risk is. The conclusion of the research provides the research direction for the future application of intelligent text algorithm to identify the default information in the description-text related borrowing.

Key words: P2P lending, description-text of borrower, default risk, willingness to repay, emotion characteristics