基于时间加权SVM的指数优化复制模型与实证分析

本文研究了指数型基金管理中最核心的指数追踪问题.依据结构风险最小化思想,首先构造了以追踪误差为基础的损失函数,接着考虑了时间因素对历史追踪误差的影响,采用指数加权的方法将时间因素纳入追踪误差的分析中,利用SVM理论给出了指数复制问题的结构风险的形式,在此基础上构建了各种市场实际约束下最小化结构风险的时间加权SVM的指数优化复制模型,最后利用OR-Library中的5个市场指数历史数据进行了实证检验.结果表明本文提出的时间加权SVM模型能够显著提高样本外的追踪效果.同时使得追踪组合的鲁棒性有明显的改善,从而表明本文提出的时间加权SVM指数复制模型具有较高的理论和应用价值.

An stock index replicating model based on time weighted SVM and it’s empirical analysis

In this paper, we study the index tracking problem which plays core role in index fund management. According to minimizing structure risk, a loss function based on tracking error is constructed firstly, then time is considered in history tracking error, and exponential weighted method is used to handle it, so the structure risk is proposed on the basis of SVM, a model is generated to deal with index tracking problem by minimizing structure risk constrained some real market conditions, lastly, some empirical tests are conducted using five index historical data from OR-Library. The numerical results show that our model can improve the out-of-sample performance, at the same time, the robust of the tracking portfolio is improved evidently, thus it has good theoretical and applied value.