| White sugar plays a vital role in our daily lives and is an essential raw material for the beverage,food and pharmaceutical industries.In recent years,the area planted with sugar crops in China has been decreasing year by year,which,together with the influence of uncertainties such as climate,has led to frequent fluctuations in the production of white sugar.According to statistics,the demand for white sugar in China has always been much higher than the production,with production not increasing and demand not decreasing,resulting in an outstanding contradiction between supply and demand for domestic white sugar.From an international perspective,China,as an important consumer of white sugar,has not had a say in the pricing of white sugar futures,and China’s white sugar does not have export competitiveness,which has led to the interests of relevant enterprises and farmers in China not being protected,and the development of the industry not being promoted,affecting the long-term stable development of China’s white sugar industry.Therefore,this paper focuses on the application research of sugar price forecasting model,which can help the relevant subjects in the sugar industry to understand the trend of sugar price changes in advance and make timely responses to it to a certain extent.The application of the futures market to the spot market is gradually expanding.The price of sugar futures is a response to the price changes in the sugar market,and the function of sugar futures hedging is of positive significance to the relevant sugar industry players in coping with the fluctuations of sugar prices.The factors influencing the price of white sugar futures are the common result of the role of various parties.Based on the background and needs of the times,it is important to pay attention to the research related to the forecast of white sugar futures prices.Therefore,this paper will consider various influencing factors and construct a prediction model for the price of white sugar futures,so as to provide reference for the relevant industries to forecast the direction of white sugar prices.At present,there are two major categories regarding the application of price forecasting methods,namely econometric models and non-econometric models.With the in-depth research on price forecasting models by scholars at home and abroad and the rapid development of artificial intelligence,the application of artificial neural networks in non-metric economic models is becoming more and more widespread and has been affirmed by the majority of scholars in price forecasting.Deep learning extended from artificial neural networks in the field of machine learning has been successfully applied to financial price forecasting,computer vision and speech recognition,natural language processing,machine translation,sentiment analysis and other fields,among which back propagation neural networks have been widely used in the analysis and prediction of financial data and have an irreplaceable role in the integration of big data related aspects.In terms of input variables,the use of non-financial data,such as climate data,is relatively rare in agricultural price forecasting studies,especially in studies on sugar futures forecasting,where climate data from all major production areas are not used as input variables to build forecasting models.Therefore,this paper will explore the factors influencing the price of white sugar,categorising and discussing the supply and demand,climate of the main production areas,international and macroeconomic aspects of the white sugar market,and finally selecting 16 relevant influencing factors to build a neural network forecasting model.The daily data from 2006 to 2020 were selected as the training data of the neural network,and the daily data from 2021 were used as the prediction data of the neural network.The final MAPE of the model was 4.03%,which proved that the model has good prediction ability.To further improve the accuracy of the prediction model,this paper will apply the reverse elimination method to explore the best combination of influence factors of the model and analyse the relevant influence factors,and take the combination with the lowest MAPE as the final input variable of the neural network.The MAPE accuracy of the optimised sugar futures forecasting model was evaluated on a daily basis for 2021,and the daily MAPE results were used to explore which time period the model is more accurate in forecasting sugar futures prices.The analysis found that the daily MAPEs for June to September 2021 generally fall in the0-5% range,which is better than the other months.In summary,this neural network forecasting model can provide the general trend of sugar price movements for relevant industrial players and has certain reference value.This study mainly focuses on the forecasting method of the model,expecting that the study can provide relevant enterprises and farmers with the direction of price forecasting,pay attention to the price fluctuation of white sugar futures in advance,and effectively avoid the risk brought by the fluctuation of white sugar prices,and at the end of the study,four suggestions are provided for the relevant government departments,enterprises and researchers as reference.Firstly,to innovate financial models for related agricultural products;secondly,to focus on the movement of the China Sugar Index in the price forecasts;thirdly,to consider the inclusion of climate factors in agricultural futures price forecasts;and fourthly,to strengthen futures education and training for related agricultural subjects.At the same time,more data and impact factors can be considered on this basis in future studies,and perhaps the construction of forecasting models by region and month can be explored.It is expected that this study will provide new ideas for sugar-related research within the sector,provide reference for the industry-related departments to forecast sugar prices,jointly maintain the stability of sugar market prices and help the healthy development of China’s sugar industry. |
PDF Full Text Request