Research On Risk Warning Of Grain Spot Trade Based On LSTM Model

Grain spot transaction has the characteristics of large transaction quantity,large price fluctuation and large transaction risk.In recent years,China's food e-commerce market has highlighted the typical characteristics of "multiple correlation of subjects,diversified trading patterns,complex and changeable risks".But,the traditional regulatory model for a single regulatory model,vertical regulation model and the single service mode,due to the difference of data sharing between each platform,associated collaborative is poor,poor coordination,model size diversity is poor,cause food spot market transaction credibility is low,grain spot transaction both sides main body inspection,food supervision spot trading process reliability is low,Grain spot trade risk early warning accuracy is bad and so on.This paper firstly summarized and analyzed the domestic and foreign status quo of grain spot trading early warning,trading early warning and the development trend of cyclic neural network,and then analyzed the difficult problems urgently needed to be solved in the current grain spot trading risk early warning.There are deficiencies in supervision and law enforcement,unclear related indicators,imperfect system,unclear objective indicators of risk early warning,and too large subjective indicators.The adaptability of the method selected by the early warning model to grain risk information needs to be further improved.Two kinds of risk warning models of grain spot trade are put forward.And the corresponding food spot trading market risk early warning workflow.Are respectively the improved algorithm Ganopso model of BP neural network and the improved model LSTM model of RNN(cyclic neural network).When applied to the risk early warning of food spot trading market,Ganopso model and LSTM model have local convergence limitations,which may easily lead to training failure in the training process.Therefore,Ganopso model and LSTM model need to optimize the algorithm so as to obtain valuable early warning results in grain spot trading.Finally,a new grain spot trading risk warning model is put forward.By using the optimized LSTM model based on deep learning,the LSTM model is easy to fall into the local optimal defect,and the warning accuracy of the model can be improved to a certain extent.In this paper,LSTM is applied to the grain spot trading problem.Based on the traditional LSTM classification model,an improved LSTM classification model is obtained by optimizing the classification model.The improved model makes full use of the input information and reduces unnecessary computation.The output of the LSTM unit model can filter out the interference information and retain the important information,so it has better classification effect and faster training speed.Experimental results show that,compared with the improved LSTM network classification model,the improved LSTM network classification model proposed in this paper has significantly improved the recognition accuracy and training speed.