In this paper, we study the issue of constructing spot portfolio which has better linkage with stock index futures. Two-stages optimization strategy is applied to improve the portfolio' s tracking accuracy. In the first stage, time series clustering method based on independent component analysis and fuzzy C-means algorithm is u-tilized and the composite stocks corresponding to the CSI 300 stock index futures are clustered. In the second stage, index optimizing replication is done with the clustering results in order to minimize the tracking error and determine the weights of tracking portfolio' s composite stocks. The empirical study indicates that the two-stages optimization strategy proposed in the paper can better improve index tracking effect.%针对如何构建与股指期货联动性较好的现货组合问题,本文提出采用两阶段优化策略以提高组合的跟踪准确度.第一阶段,利用基于独立成分分析与模糊C均值算法相结合的时间序列聚类方法将沪深300股指期货对应的成分股进行聚类；第二阶段,对聚类之后的结果进行指数优化复制,以跟踪误差最小为目标,确定跟踪 组合的成分股权重.实证研究表明,本文所提出的两阶段优化策略可以较好地改进指数跟踪效果.
展开▼
