Price Forecast Of Light Rare Earth Products Considering Supply And Demand,Finance And Phase Characteristics

Rare earth is a very important element with rich properties,especially in high-tech fields.China's light rare earth reserves are abundant,mainly concentrated in the Baiyun Obo mining area of Baotou City.Its reserves account for more than 90% of the total rare earth reserves in the country,ranking first in the world.However,the price of light rare earth is low and the price fluctuates greatly.Especially in 2011,the price fluctuated drastically,and the rare earth resources in the process of mining and smelting made China pay a huge environmental cost.In the Sino-US trade war in 2018,rare earth as a potential trump card has also received great attention.Nowadays,economic and trade risks and environmental pollution problems are becoming more and more prominent.Price forecasting of rare earth products is not only an important task to ensure the healthy and sustainable development of countries,industries and enterprises,but also a necessary means for effectively utilizing rare earth resources,reducing environmental pollution and avoiding the risk of price fluctuations.In this paper,praseodymium oxide,neodymium oxide and cerium oxide,the most widely used rare earth products in the downstream market of light rare earth permanent magnets,are studied.This paper not only considers the domestic and international downstream industry demand factors,but also considers the financial factors that have a weak direct impact on the spot market price of rare earth products or indirect impact on the price of rare earth products through the influence of other commodities.The nonlinear research method was used to construct the model of cerium oxide,neodymium oxide and praseodymium oxide considering the financial factors and demand factors.The MSVAR model of cerium oxide,cerium oxide and cerium oxide with regional conversion characteristics were obtained.The results show that all three products show three zones: regime 1 shows downward fluctuation characteristics,regime 2 shows stable fluctuation characteristics,regime 3 shows upward fluctuation characteristics;There are significant differences between domestic and foreign downstream industries in the impact of three product regional on prices.The negative relationship between LRE products and downstream industries reflects the active demand management ability of downstream industries.The positive relationship between LRE product price growth rate and downstream industries does not reflect the demand management ability of downstream industries.Through the above-mentioned non-linear model MSVAR,we find out the significant influencing factors of three products under different regional conditions,and bring them into the decision tree and combination forecasting(Bagging,AdaBoost and random forest)forecasting model,and compare the forecasting results with the actual values to judge the forecasting effect of different forecasting methods.The results show that,from the point of view of prediction methods,the prediction effect of adabag is the best.From the point of view of regional system,the prediction effect of regime 3 of three products is better than that of regime 1 and regime 2.The forecasting results can provide some reference information for enterprises in the light rare earth industry chain in making decision to reduce the risk of price fluctuation of light rare earth products.