Research On Futures Trading Strategy Based On Complex Network Factors

As China's futures market system continues to enrich and improve,and as investor styles change,the futures market as a complex system is in a constant state of change.This changing state of affairs has led to new requirements for the construction of factors related to futures portfolios and trading strategies that reflect the nature and characteristics of the market in a timely and high-quality manner.In the face of this market demand,it is important that the system of relevant factors for futures markets is constantly updated and developed in line with the new environment and technology of the day,in order to provide a more effective portrayal of the characteristics of futures markets while matching the increasingly diverse investment preferences of futures investors as far as possible.Using complex network theory,this thesis selects 54 available futures varieties in China's commodity futures market as network nodes and uses the correlation between the log returns of each futures variety's main contract as the connecting edge to construct a return correlation network for China's futures market.Next,the correlations among China's commodity futures varieties are analysed,and the overall characteristics of China's commodity futures return correlation network are analysed.The results of the characterisation of this network show that China's futures return correlation network is a typical small-world network at a small threshold.The centre of the network is occupied by the energy and chemical futures species,and the returns of this category of futures species have a greater impact on the overall futures return correlation network.By analysing the characteristics of the nodes in the futures return correlation network,three candidate complex network factors,namely weighting,clustering coefficient and eigenvector centrality,are selected,and the portfolio is constructed by ranking the corresponding factor values of the futures return correlation network on a monthly and quarterly basis over the sample period.The "Trend Tracking Double Average + Percentage Trailing Stop" trading strategy was used to obtain the returns of each portfolio for the following month and quarter,and the validity of the factors was determined by testing the significance of the series of return differences for the portfolios with different factor sizes.The next step continues with the analysis of the returns obtained in future quantitative trades using the resulting portfolios.In this thesis,the two complex network factors of weighting degree and clustering coefficient obtained from the complex network of China's commodity futures return correlation can be used to construct portfolios with significant positive returns,which not only enriches the system of correlation factors in China's futures market,but also serves as a reference and help for the design and construction of quantitative investment trading strategies in China's futures market.