Futures Correlation Analysis Based On Spatial Weight Matrix

With the continuous development of futures market and quantitative trading,the market has more and more requirements for investment decisions.How to timely and accurately grasp the development law of the market is the premise of stable profit in trading.Mathematical models and probabilistic theories are widely used in the price prediction of financial markets,as well as the ways of making investment decisions based on the relevant analysis of transaction data,which are more and more favored by financial institutions and investors.In recent years,machine learning algorithms have attracted the attention of financial market technical analysts,and are widely used in the field of financial asset prediction,which provides great help to solve the problems of financial market.Therefore,the main research content of this paper is to combine the spatial measurement model and the factor model,integrate the advantages of the two types of models,and model them through machine learning algorithm tools.The correlation of prices between varieties in the futures market is analyzed by two types of models,and the future trend of prices in China’s futures market is predicted.First of all,relevant studies show that the spatial measurement model and the factor model have outstanding performance in the estimation of weak correlation and strong correlation fields,respectively,which is very suitable for the study of the price trend in the futures market.Therefore,this paper adopts the combination of spatial measurement model,factor model and machine learning algorithm,and explores the change of the price correlation of various futures varieties in different cycles through the innovation of model and methodology.The empirical results show that the price correlation factor of a long period has a stable spatial weak correlation;the price correlation is consistent with the price trend.Secondly,this paper reasonably classifies the setting of the spatial weight matrix in the spatial measurement model,and uses it to analyze the degree of correlation between the prices of various futures varieties.In the research practice,it is found that the traditional pearson correlation coefficient method has its own advantages compared with the three types of spatial weight matrix correlation coefficient method divided in this paper,and has its own focus in exploring the correlation of futures.Finally,based on the price correlation analysis of various futures varieties,this paper predicts the short and long price trend through certain mathematical model algorithm.By predicting the changing trend of the future price,the corresponding investment strategy is provided to provide certain theoretical support for the arbitrage trade.The empirical analysis shows that the prediction accuracy of short cycle price trend is significantly higher than that of long cycle price trend;after extracting the strong correlation,the extraction strength of the weak correlation greatly affects the accuracy of the price prediction model;...