With the development of the Internet, Internet finance in new P2P modes will face a great many difficulties and opportunities; so, relevant risk early-warning models need to be researched and analyzed. The early-warning analysis will not only be helpful for P2P, the new mode, but will also be worth learning by the whole Internet financial industries, and there will be a particular demonstration effect. Deep researches have been made on Internet financial risk precautions mainly through analyzing and researching the risks in the leading P2P online debit and credit model within the scope of Internet finance; therefore, risk factors that influence the development of Internet finance are obtained. Next weighting KNN Internet financial risk management and control algorithm with the variable precision rough set is out forward. Training sets of different categories are divided into positive regions and boundary regions through the upper and lower approximation concept of variable precision rough set, thereby acquiring the affiliation regions of the samples based on the similarity between test samples and the sample center. In this way, the category of samples belonging to the positive region can be directly judged, and that of other regions can be judged through the KNN algorithm based on quantitative weighting. Experimental results have verified the effectiveness of the mentioned algorithm.

随着互联网的发展，新的P2P模式下的互联网金融将面临许多困难和机遇。因此，需要研究和分析相关的风险预警模型。预警分析不仅对新的P2P模式有所帮助，而且值得整个互联网金融业学习，并且会产生一定的示范效果。主要通过对互联网金融范围内领先的P2P在线借贷模型中的风险进行分析和研究，对互联网金融风险防范措施进行了深入研究。因此，获得了影响互联网金融发展的风险因素。提出了基于变精度粗糙集的加权KNN互联网金融风险管理控制算法。通过变精度粗糙集的上下近似概念，将不同类别的训练集分为正区域和边界区域，从而根据测试样本与样本中心之间的相似度，获取样本的隶属区域。这样，可以直接判断属于正区域的样本的类别，而通过基于定量加权的KNN算法可以判断其他区域的样本的类别。实验结果证明了所提算法的有效性。这样，可以直接判断属于正区域的样本的类别，而通过基于定量加权的KNN算法可以判断其他区域的样本的类别。实验结果证明了所提算法的有效性。这样，可以直接判断属于正区域的样本的类别，而通过基于定量加权的KNN算法可以判断其他区域的样本的类别。实验结果证明了所提算法的有效性。