A Fast Algorithm For Moving Dickey-Fuller Statistic And Its Application

This paper focuses on a fast algorithm for moving Dickey-Fuller statistic(hereinafter referred to as DF statistic)and its application.The problem stems from the fact that the traders who use the pair trading strategy want to be able to determine whether the latest quotation of the two given assets can pass the EG test,to establish the cointegration relationship and determine the investment portfolio.Although the problem is ultimately to serve the actual financial problem,this paper does not discuss the effectiveness of the investment strategy involved,but rather limits the model to the mathematical level,through a series of mathematical deduction and the use of statistical methods to significantly improve the efficiency of the calculation,to provide technical support for the relevant practitioners.In the empirical analysis stage,we select a piece of data of palm oil commodity futures and soybean oil futures to verify the effectiveness of the algorithm,and provide solutions for the problems we may encounter while applying the algorithm to the actual scene.