Research On Price Prediction Of Rare Earth Products From The Perspective Of Price Influencing Factors

Rare earth is an extremely important element with rich properties.As a strategic resource in China,it plays an important role in the field of high and new technology.The total reserves of rare earth in China ranks first in the world.The strategic position of rare earth has undoubtedly become a potential trump card in China's trade war.However,the lack of management in the rare earth market for a long time has led to the low price of rare earth,serious smuggling,and unlimited mining,which has caused environmental pollution problems,which has caused great economic losses in China,and is not conducive to mastering the pricing power of rare earth.Therefore,it is of great significance for China to master the law of price fluctuation and clarify the influencing factors of price fluctuation,which is also a necessary means to effectively avoid price fluctuation.Starting from the influencing factors of the price of rare earth resources,this paper considers the supply and demand factors and financial factors as the influencing factors of the price fluctuation of rare earth products.A BP neural network(ACO-BP)combination model based on ant colony optimization algorithm is constructed to predict rare earth products.Firstly,principal component analysis(PCA)is used to eliminate the redundant information between the influencing factors of rare earth price prediction,reduce the dimension of the input data of BP neural network,and improve the prediction accuracy;secondly,ant colony algorithm is used to find the optimal neural network threshold,so as to reduce the prediction error at the convergence speed of the optimization model.In this paper,neodymium oxide,lanthanum oxide,cerium oxide and dysprosium oxide are studied.Based on the monthly data of January 2010-2018 March,the multi factor ACO-BP combination model is constructed to forecast.Compared with the prediction results of BP neural network model without optimization,the comparison results show that the multi factor ACO-BP combination model is better than the traditional BP neural network model in simulation ability,error level,convergence accuracy,etc.,and can more accurately predict the price trend of rare earth oxidation,which has a certain guiding role for the price formulation of rare earth products,and also for other time series The prediction analysis of column data provides another feasible way.According to the prediction results,this paper constructs the early-warning system of rare earth price fluctuation to predict and forewarn the price fluctuation.According to the actual economic market operation,industry integration,national political decision-making and other factors,the system can adjust and optimize parameters,predict the fluctuation trend of rare earth price,and timely issue the fluctuation warning signal.Relevant decision makers can make timely adjustment of relevant policies according to early warning signals,so as to reduce market risk.We will set up a three-level early warning system,namely,the first level risk early warning,the second level risk early warning and the third level risk early warning.Set three confidence intervals,i.e.the corresponding confidence intervals with confidence levels of 0.99,0.95 and 0.9,and divide the rise and fall of predicted values into intervals.Assist the government and management department to specify corresponding policies to reduce the risk of price fluctuation.