Research On Price Forecasting Of Gannan Navel Based On BP Neural Network

Gannan navel is an important fruit of citrus in China,ranking first in the agri cultural region of the public brand value in 2105.And it played a very important role in improving the brand awareness of the Gannan tourism,driving the people t o get rid of poverty,promoting the transformation and upgrading of Gannan navel industry.Before 2016,Gannan navel's planting scale more than 1 million 600 thousands acres,industrial output value nearly 10 billions,Planting,processing,sales gradual-ly increased,the consumption of Gannan nave increased year by year in the fruit market,it's market size gradual expanded,however,Gannan navel is also facing fie rce price fluctuations and short term volatility,it may harmful to consumers when the price is high,and harmful to grower when the price is low.The stability of Gannan navel's prices have important relationship with the income and life stabilit y of producers,processors and consumers.Study on forecasting of Gannan navel's price is benefit to formulating market policy and sales strategy,it is of great signi ficance to guarantee the stable of farmers' income.The paper is based on the theory of price formation,Analysis the fluctuation characteristics and influencing factors according to 2007-2017 monthly price data and the daily price data from November 2016 to February 2017 of Gannan navel,Forecasting the price of Gannan navel based on BP neural network,and then compared the prediction accuracy of BP neural network and ARIMA model,according the results,analyses were conducted on the price trend of Gannan navel,Made empirical analysis using MTLAB software.Therefore,the research contents and conclusions are as follows:(1)Study on the industry development and price changes of Gannan navel;Analysis the problem of Gannan navel from the price fluctuation,industrial structure and market promotion;Discussing the main factors that affect the price of Gannan navel from three aspects of market factors,natural factors and human factors.(2)Forecasting the price of Gannan navel.Determine the number of nodes in hidden layer according to thecharacteristics of Gannan navel,built BP neural network price forecasting model based on improved L-M algorithm.The prediction error of BP neural network is lower,and the relative error is 2.5%.Compared with the ARIMA model,the relative error of BP neural network prediction is 1.5%,lower than the relative error of ARIMA model which is 2%.(3)Providing countermeasures about the fluctuation of the price of Gannan navel.(3)Based on the result of price forecast and the cause of fluctuation,we think the Gannan area should regards the construction of information,strengthen the early warning of price fluctuation of Gannan navel,improve the scale of operation,get the production cost lower,improve the resilience of price volatility risk.The innovation points of this paper are as follows:(1)In the field of research,selecting the latest price data,for reducing the impact of the long time span.In addition,there are a lot of researches on the price forecasting of fruit,but there are few studies on the price prediction of Gannan navel.We applied the intelligent forecasting method to the area of Gannan navel which get rapid development in recent years,it is very different from the previous qualitative analysis.(2)About the selection of prediction methods,comparing the goodness of standard BP neural network with the improved L-M algorithm,compared the BP neural network with ARIMA model in prediction accuracy,selected appropriate forecasting model to avoid the model simplification.