基于多元线性回归的上证50股指追踪研究

Research on Tracking of Shanghai 50 Stock Index Based on Multiple Linear Regression

摘要: 指数追踪是指通过利用一个的股票组合复制某一现实指数或者虚拟指数的市场表现，由此来得到目标指数的市场表现，并尝试最小化跟踪误差。其目的是追踪一个股票指数的持仓及盈利表现。本文在追踪之前，对数据进行了回归诊断，诊断结果表明变量间存在多重共线性，逐步回归与岭回归方法能够很好的消除多重共线性。因此，本文主要采用了逐步回归与岭回归对上证50指数的5分钟K线数据进行指数追踪。指数追踪结果表明，利用逐步回归法对上证50指数的追踪效果优于岭回归法。

Abstract: Index tracking refers to using a stock portfolio to replicate the market performance of a real or virtual index to obtain the market performance of the target index and try to minimize tracking errors. Its purpose is to track the holdings and earnings performance of a stock index. Before tracking, this paper carried out regression diagnosis on the data. The diagnosis results showed that there was multicollinearity among the variables. Stepwise regression and ridge regression methods can eliminate multicollinearity very well. Therefore, this paper mainly uses stepwise regression and ridge regression to track the 5-minute K-line data of the Shanghai Stock Exchange 50 Index. The index tracking results show that the tracking effect of the SSE 50 index using the stepwise regression method is better than that of the ridge regression method.