基于Adaptive Lasso的股票指数跟踪问题研究

Research of Stock Index Tracking Based on Adaptive Lasso

将处理高维数变量选择问题的Adaptive Lasso运用于指数跟踪问题,利用LQA算法,实现Adaptive Lasso的求解,为构建股票投资组合跟踪指数提供了一种新的方法。实证部分表明:利用此方法构建股票组合来跟踪沪深300指数能够获得很小的跟踪误差,预测能力十分可观,且计算时间极短,说明了该方法在指数跟踪中的实用性。

Adaptive Lasso,which is often used in variable selection on high-dimension data was used for index tracking,and LQA algorithm was adopted to solve Adaptive Lasso problem,thus providing a new approach to build a portfolio tracking index.The empirical results show that using this method to build portfolio to track the CSI 300 index can obtain a small tracking error with impressive predictive ability and very short calculation time,thus proving the practicability of this method.