Research On Price Prediction Of Agricultural Products In Some Areas Of Zhejiang Province Based On Model Analysis

China is one of the most important agricultural countries in the world,and its rich agricultural resources provide important support for people’s daily needs,while also contributing to the country’s economic growth.The change of agricultural price is not only related to the interests of agricultural producers,sellers and consumers,but also related to the stable development of the country.Zhejiang province as a large trade province of agricultural products,has rich agricultural resources,many agricultural products in the production of the country have obvious advantages.Studying the price of agricultural products in Zhejiang province and being able to predict it more accurately is also of great guiding significance to the national agricultural industry.Based on this purpose,this thesis has done the following work:First of all,this thesis analyzed the factors affecting the price of agricultural products by collecting and sorting out the researches of many scholars on the factors affecting the price of agricultural products.The price factors can be roughly divided into the following nine aspects: resident income level,substitute prices,consumption habits,emergencies,output,production costs,market circulation,government policies and natural factors.Taking tomatoes in Ningbo city and mushrooms in Jiaxing city as examples,relevant factors affecting agricultural prices were collected according to the availability and authenticity of influencing factor data,and correlation analysis and principal component analysis were carried out to lay the foundation for subsequent research.Secondly,according to the obvious seasonality of the price changes of agricultural products,the seasonal decomposition forecasting method is selected to forecast the seasonal variation components of agricultural products prices.Based on the influence of historical agricultural prices on existing prices,exponential smoothing method and ARIMA(differential autoregressive moving average model)are selected.According to the effect of various factors on the price of agricultural products,a multiple linear regression model is selected.The price of tomato and mushroom were predicted by the above four models,and the results of the models were evaluated by MAPE(mean absolute percentage error).Compared and analyzed the prediction results of four models,the model effect of ARIMA is the best,followed by exponential smoothing method,season decomposition prediction method and multiple linear regression model.Therefore,for a single forecasting model,better results can be obtained by using ARIMA model.Finally,on the basis of building a single model,this thesis proposes two combination models,one is the combination model of seasonal decomposition forecasting method and exponential smoothing method,the other is the combination model of ARIMA and multiple linear regression.Comparing and analyzing the prediction results of all models,it is concluded that the prediction result of the combined model is better than that of any single model,and the prediction effect of the combined model of seasonal decomposition forecasting method and exponential smoothing method is better than that of the combined model of ARIMA and multiple linear regression,which can realize the effective prediction to the future price of agricultural products.