Volatility And Short-term Prediction Of Litchi Price In China

Recently,price experienced more and more intense and frequent volatility initchi market in C hina.Litchi price fluctuationsaffected the income of litchi producers,the welfare of Chinese litchi consumers,and the profit of other stakeholders in the supply chain of litchi industry.Therefore,the price fluctuation of litchi market is not only an economic problem,but also a kind of interest from producer to consumer in litchi supply chain.So,it is of practical significance to analyze the characteristics of litchi price fluctuation and make short-term forecast.The main contents and results of this dissertation are as follows:First,this dissertation combedthe internal transmission mechanism and external shock factors influencing the price fluctuation of agro-products,eight characteristics of agroproduct price fluctuation,and the research method s of price fluctuation of agro-products.Based on the existing research results,Eviews8.0 software and STATA14 were applied to realize BP filter and HP filter method s.And then the dissertation analyzedlitchi short-term fluctuation characteristics of the purchase price in originsand the wholesale price insales territories,and summarized the fluctuation characteristics ofprice temporal transmission in origins and pins,respectively.After BP filter and HP filter oflitchi purchase price and wholesale price,the main conclusions are as followes:(1)Both litchi purchase price and wholesale price are strong cyclical;(2)The volatility periods of litchi purchase price and wholesale price fluctuations are unrepeated;(3)Fluctuations of litchi purchase price and wholesale price complied with asymmetric transmission;(4)The price fluctuation of litchiwholesale price is consistent with the characteristics of "rocket and feather" from previous studies,but the new conclusion of this dissertation is that litchipurchase price volatility is anti-"rocket and feather" asymmetric characteristics.The previous three conclusions are consistent with those of the existing literature,and the last one is a new discovery of this dissertation.In addition,another innovation of this dissertation is that the research perspective of litchi price fluctuation is different from that of vertical transmission or spatial transmission.In this dissertation,temporal transmission and spatialtransmissionwere combined together,but temporal transmission was more focused on..In addition,this dissertation analyzed the reasons of the four fluctuation characteristics of litchi prices--the periodicity,the non-repetition periodicity,the asymmetry and the different asymmetric transmission in origins and sales areas.Second,this dissertationused MATLAB software to realize BP Neural Network,R software to realize Support Vector Machine,and LINGO12 software to realize Composite Modelrespectively,and analyzed the direction,amplitude and frequency of volatility of the litchi price to predict the coming 60 daily purchase prices and wholesale prices in 2017 of 10 originsand 11 sales areas.The empirical results show that BP Neural Network,Support Vector Machine and Composite Model have highforecast accuracy for purchase price and wholesale price,and the prediction res ults of the three models are basically the same.This dissertation drew the following conclusions: in the future,the purchase prices in litchi origins and the wholesale prices of sales areas will have convergence effect of fluctuation trends,fluctuation ranges and fluctuation frequencies.Finally,this dissertation pointed out thepossible research directionsin the future.