Research on Strategies of Stock Index Futures Transaction Based on LMS Linear Regression

Abstract

In recent years, stock index futures has developed into an important financial derivative in the global capital market and provided an indispensable investment and hedging tool for investors. This paper is mainly studying the low frequency transaction of stock index futures. Firstly, it conducts data cleaning to the stock index futures data and has an interpolation calculation to the missing values. Afterwards, it judges the market information in the short period and achieves the purpose of gaining profits through the basis. Furthermore, the time and the price difference are used as the key factors for establishing the quantitative model of gaining profits. Through lease square regression and inventory analysis technology, it completes the whole transaction process with several actions such as judging the open position time, transaction direction, reverse position closing time and yield accounting. Finally, as to the regression testing of the model, it detects the expression of the strategy through annual earnings, maximum drawback, transaction frequency, Sharpe ratio, variety commonality and cycle commonality. In the end, it conducts scientific analysis to the approaches used in the model and discusses the advantages and disadvantages of the model, thus providing some optimizing strategies in considering the improvement direction in practical application.