Using Weighted Similarity to Assess Risk of Illegal Fund Raising in Online P2P Lending

Abstract

Peer-to-peer (P2P) lending is an important internet financial mode, which has a greater risk of illegal fund raising. From the risk research on P2P lending platforms has focused on policy and law, and the existing risk assessment is mainly aimed at borrowers' credit. Since it cannot meet the needs of effective supervision, this article proposes a risk alarm model from the perspective of illegal fund raising based on similarity weighted case. Through the investigation of P2P illegal fundraising cases, this article has extracted the risk features to build a risk feature matrix. A case to be evaluated needs to be transformed into a feature vector in the data preprocessing stage. Then, the similarity vector can be obtained by comparing a feature vector with the vectors in the risk feature matrix. The following selected the TOP K similarity to calculate the risk value by weighting. The experiments show that under the condition of even a small sample, it can reasonably evaluate the risk of the P2P lending platform, to achieve a certain risk alarm effect, and has a good feasibility.