Research On Quantitative Investment Strategy Of Commodity Futures Based On Random Forest

This paper constructs a quantitative investment strategy for China's commodity futures market,and uses the volatility classification of random forests to filter the entry conditions of the strategy.As a result,the performance of the strategy has been greatly improved.In the end,34 well-fluid futures varieties in the commodity futures market were used to construct the investment portfolio.Through the comparison of the internal and external strategies of the sample,it was verified that the random forest machine learning tool can improve the effectiveness of the quantitative strategy in China's commodity futures market.In the process of constructing a quantitative strategy,this paper based on the idea of interval breakthrough,and classified the market volatility to improve the strategy.Finally,the historical data was used to verify the effectiveness of this improvement.In the process of classification,this paper selects a machine learning tool,random forest,to classify the volatility of market conditions nonlinearly.This classification improves the ability of the range breakout strategy to adapt to various varieties and markets in commodity futures investment,thereby improving signal quality.At present,with the widespread adoption of computer technology and data mining tools in the financial industry,the investment concept of quantitative investment is increasingly used in secondary market investment.Among the many kinds of quantitative investment strategies,the commodity futures investment consultancy strategy is one of the mainstream ways to quantify investment in commodity futures.The strategy focuses on the volatility of commodity prices,using quantitative strategies to give investment advice,and eventually gaining positive returns by going long or shorting a basket of commodity combinations.With the rise of machine learning and artificial intelligence,many domestic and foreign scholars have tried to apply machine learning tools to the analysis of the securities market.Although there have also been some attempts to apply machine learning tools to quantitative investment strategies in the financial market,However,there are no relevant concrete results.Random forest is an ensemble learning method in machine learning.Compared with other tools,it has many advantages,such as: it is not easy to overfit and has higher tolerance to noise.Therefore,it is particularly suitable for quantitative investment strategies.Most of the current market breakout strategies are using traditional models to judge the direction of trends.There is no precedent for combining machine learning.Can machine learning be used in quantitative investment? This paper hopes to provide a good example for the combination of CTA strategy and machine learning through random machine learning tool,and also provides investors with a new research idea.