Research On Agricultural Product Price Forecast Based On Data Mining

The market price of agricultural products is related to people's livelihood and social stability.The fluctuating prices of agricultural product prices will affect consumer demand.The processing industry with agricultural products as raw materials will easily cause people to panic in frequent price fluctuations,thus accelerating the vicious circle of price fluctuations.In recent years,due to the lack of authoritative information,there have been many phenomena of daily consumption of agricultural products that have skyrocketed and plummeted.The price of agricultural products is affected by a variety of natural social factors.Natural disasters,supplies,and demands for substitutes for agricultural products can cause violent fluctuations in agricultural product prices,which increase the difficulty for government departments to supervise market of agricultural products.In order to promote the stable development of the agricultural market and increase the income of farmers,the government departments have repeatedly emphasized the importance of supervising the agricultural products market and issued important report documents.Therefore,the development of short-term forecasting of agricultural product prices has important theoretical and practical significance for supervising agricultural product market prices,stabilizing agricultural production demand,and ensuring people's living standards.In this dissertation,the prices of pork and garlic were selected as the research object.And the price index and its influencing factors for more than ten years were analyzed.Guided by the related ideas of economics,market finance,statistics,etc.,combined with Chinese agricultural product price policy,BP neural network,RBF neural network,NARX neural network and other research methods were used to explore the price of agricultural products in China,which investigates the fluctuation law of Chinese agricultural product price under the interweaving of various influencing factors,breaks the traditional statistical forecasting method of traditional agricultural product price,and establishes the short-term prediction model of agricultural product market price with innovative technology.The steps of agricultural product price index analysis were described as follows:Firstly,the price influencing factors were analyzed mathematically,all data were preprocessed and the principal components were extracted,and works were prepared for the neural network experiment.Secondly,an experimental model was structured under the guidance of data mining ideas.In this dissertation,three experimental models were established for agricultural product price prediction: BP neural network model,RBF neural network model and NARX neural network model.After statistically summarizing the influencing factors,the number of optimal hidden layer nodes were selected by setting formulas and experiments.By adjusting the transfer function and values,the optimal neural network models were obtained.The network model was constructed to predict the agricultural product price index.Finally,the prediction values of the three network models were calculated and compared with the overall error between the measured values.It is concluded that the BP neural network model and the NARX neural network model have excellent accuracy in different agricultural product price prediction indexes.The data mining has the practical value and the vital significance to agricultural product price market analysis and decisionmaking.