Decision Support System Of Agricultural Science Development Strategy Based On Data Mining

As an important part of China's national economic system, agriculture is the basic product of national construction and development. It is an important lifeblood of the national economy and a strong guarantee of national long-term stability and well-being of the people. With the rapid development of economy and society, people has produced a large number of agricultural data and information, including agricultural prices. People are no longer satisfied with the superficial data manipulation, hoping to find valuable knowledge and market information from the massive data.In recent years, the price of agricultural products fluctuates frequently and presents unstable trend, which directly affects the healthy development of agriculture and the normal life of the people. In this paper, we propose a scientific method based on time series data analysis to forecast agricultural prices, considering the weakness of China's agricultural theory foundation, the lack of information arrangement and short-term price forecast research. The main idea of time series analysis based on data mining is to quantitatively describe the law of price variation with mathematical models and judge the future value by historical price values.We deeply studied the classic ARIMA model, and established the ARIMA(5, 1, 6) to analyze the cantaloupe price in Beijing Xinfadi farm produce wholesale market and observe the ARIMA model performance.The results show that the ARIMA model has a good forecasting performance, and it can provide effective and reliable information for agricultural practitioners. Furthermore, this paper investigates the economic factors that influence the price fluctuation of agricultural products, adding the CPI of the Beijing consumer price index to the ARIMA model and form the error correction model. The results show that the error correction model is more accurate than the ARIMA model, having a better performance to reflect the fluctuation of the monthly cantaloupe price. In the process of modelling, we found that the cantaloupe price is 12.95% positively correlated with CPI by rising 1 percent in Xin Fadi Wholesale Market, which can help local people understand the price fluctuation of local cantaloupe and make a timely and reasonable schedule according to the CPI. Meanwhile, we designed a practical agricultural price forecasting system. It helps not only to grasp the market development dynamics by providing meaningful agricultural information service, but also to have a contribution to making correct market decisions for agricultural practitioners.