本文研究了指数型基金管理和指数套利中最核心的指数追踪问题。依据结构风险最小化思想,建立了基于支持向量机的指数追踪模型,并利用OR-Library中的测试数据之恒生指数历史数据进行了实证检验。数据结果表明本文提出的新方法能够提高样本外的追踪效果,同时也说明该方法具有良好的鲁棒性,从而具有较高的理论和实用价值。 This article examines the core index tracking of index fund management and index arbitrage. According to the concept of minimization of structural risk, an exponential tracking model based on support vector machine is established, and the empirical test is conducted by using the historical data of Hang Seng Index of the test data in OR-Library. The data results show that the new method proposed in this paper can improve the tracking performance outside the sample, and also shows that the method has good robustness, which has a high theoretical and practical value.