摘要
场外配资是一种高风险的融资手段,对可疑的场外配资行为进行及时的识别与监控,有利于维护投资者的合法权益与证券市场的稳定。为此提出一种基于改进XGBoost机器学习算法的配资账户识别方法。通过分析场外配资的业务逻辑,构建了与识别算法强相关的特征指标体系,并结合场外配资行为特性采用召回率作为关键度量指标。通过对所构建识别算法的对比分析,所提出的基于XGBoost的场外配资识别模型得到了更加准确的识别效果,并且通过市场交易行为分析能够更加灵活快速适应市场环境变化,从而更好地用于证券市场的场外配资监控。
Shadow margin financing is a high-risk financial financing tool,and timely detection of suspicious financing behavior is conducive to maintaining the interests of users and the stable operation of the securities market.This paper proposes a novel approach to identifying the funding account based on an improved XGBoost machine learning algorithm.By analyzing the business logic of shadow margin financing,a new feature with strong correlation recognition algorithm is constructed.Recall is used as the measurement index to compare the recognition effect under different test set and training set proportions.Through the comparative analysis of the models,it is found that the overall effect of the improved XGBoost model based on binary classification is better,which can be better used in the actual business of shadow margin financing monitoring.
作者
俞建群
李双宏
YU Jianqun;LI Shuanghong(Orient Securities Company Limited, Shanghai 200010)
出处
《微型电脑应用》
2021年第12期33-36,47,共5页
Microcomputer Applications
基金
上海市信息化发展专项资金(信息化建设和应用)资助项目(201901060)。
