Prediction Model Research Of Apple Price Based On Big Data

Apple has high nutritional value and is distributed all over the world.The apple production of China has accounted for more than half of the world's total output.Through the research and analysis of previous apple price data,accurate predictions of future apple prices are achieved,which can help fruit farmers and relevant personnel in the apple industry to adjust their planting and sales strategies,and alleviate the adverse effects of price fluctuations on the apple industry.It is of great significance to the healthy development of apple industry.This article aims to construct an apple price change model,predict future apple prices in the short term and average apple prices in the coming year,and provide a reference for relevant departments,apple growers,and consumers to grasp the changes in apple prices.Firstly,the web crawler is used to capture real-time apple price data from the agricultural product service platform of the Ministry of Commerce.By analyzing the characteristics of the obtained data,the data is preprocessed by using data cleaning and data transformation.Secondly,by comparing the advantages and disadvantages of various domestic and foreign agricultural product price prediction methods,the methods of ARIMA and exponential smoothing are selected to construct model.Furthermore,in order to improve the prediction accuracy,combining the characteristics of ARIMA method and exponential smoothing method,a linear model based on ARIMA method and one-time exponential smoothing method is constructed,and genetic algorithm is adopted to select the weight of two single models in the combined model.Finally,apple price data of 2019 is used to verify the ARIMA model,the exponential smoothing model,and the combined model.The average absolute percentage error of the ARIMA model is 4.08%,the average absolute percentage error of the first exponential smoothing model is 4.60%,and the average absolute percentage error of the combined model is3.26%.These results show that the combined model has the smallest prediction error and the highest accuracy.The combined model of apple price constructed in this paper can predict the changes of apple price in the short term and the next year.The modeling process can provide reference to build other fruit price prediction models with the same or similar data characteristics.