Design And Research Of Agricultural Product Price Prediction And Analysis System Based On Hadoop Ecosystem

Agriculture is the foundation of our national economy.At present,there is a certain degree of disconnection between Chinese agricultural production and the market.It needs a scientific and reasonable basis to promote the effective formulation of agricultural production decision-making.Based on big data,taking the price information of agricultural products as the subject,this paper forecasts the price trend of agricultural products,analyzes the influencing factors of agricultural products price,and looks for a big data analysis and processing method more suitable for Chinese agricultural production decision-making,in order to use the existing big data technology to improve the scientificity of agricultural production decision-making.Based on big data,this paper designs and implements the prediction and analysis process of agricultural product prices.Firstly,it uses the web crawler tool to crawl the price information of agricultural products on Chinese agricultural websites.Then,it uses Hadoop ecological related technology to build a big data platform for agricultural products price,which overcomes the shortage of massive data storage of traditional technology.Finally,it uses the time series prediction algorithm to predict and analyze the price of agricultural products.The agricultural product price big data platforms built in this paper are implemented by using open source framework,in which Python crawler technology is selected for big data collection and Hadoop is selected for big data platform.Taking HDFS,HBase and hive as storage systems.Taking Map Reduce as a parallel offline computing framework and using sqoop to realize the data import and export from Mysql to big data platform.The algorithm model is used to compare and predict the pork price,analyze the influencing factors of pork price fluctuation,and build a pork price prediction and analysis process suitable for big data scenarios.Through the comparison and verification of the prediction results of Second exponential smoothing method and Holt-Winters method,it is found that the root mean square error and average percentage error of the prediction results of Holt-Winters method are smaller,and the prediction effect is more significant.Finally,Holt-Winters method is used to predict and analyze the pork price trend,and the pork price prediction information is obtained.This information can provide a basis for the decision-making of pig breeding.