基于支持向量机的私募股权投资风险预测

由于私募股权基金在运作过程中存在着多层委托代理关系，从而产生了不同利益主体之间的信息不对称，并进而造成了逆向选择和道德风险问题，这是私募股权投资产生风险的最主要根源，这种投资风险可以从宏观和微观两个层面进行识别和预测。基于此，可构建私募股权投资风险的备选指标体系和支持向量机模型，并进行实证评价。实证结果表明：私募股权投资风险随着投资周期的增加而提高，相应的投资风险也因各种不确定因素的增加而呈现上升趋势；SVM方法对私募股权投资风险能够进行有效预测，这将为私募股权投资风险的预测提供理论指导和方法借鉴。

Prediction of Investment Risk of Private Equity Based on Support Vector Machine

The multi-level principal-agent relationships in the operation of private equity fund produce information asymmetry among different stakeholders.Adverse selection and moral hazard problems are therefore caused,which is the most important source of private equity investment risks.The paper claims that the investment risks can be identified and predicted from macro and micro levels.Based on this assumption,this research constructs an alternative index system and support vector machine model for private equity investment risk and evaluates them with empirical evidence.The results show that,first,the risk of private equity investment increases with the increase of the investment cycle, and the corresponding investment risk rises along with the increase of various indeterminate factors;second,the SVM method can effectively predict the risk of private equity investment,which will pro-vide theoretical guidance and methods for predicting the investment risk of private equity.