Analysis And Research Of Medicine Cold Chain Data Based On Circulating Neural Network

The safety and effectiveness of drugs are closely related to the environmental variables in which the drugs are located.Because the storage time of drugs is relatively short and vulnerable to the surrounding environment,in order to solve the problem of drug quality in all cold chain links of storage and transportation of low-temperature drugs,accurate prediction Temperature and humidity are essential for the storage of medicines in the pharmaceutical cold chain.In this paper,aiming at the regulation of temperature and humidity in the quality management of medical cold chain,a zero-stock warehouse and pharmacy of a medical distribution center is taken as the research object,and a long-term short-term memory(LSTM)based on improvement is proposed.The method of predicting the temperature and humidity data of medical cold chain in circulating neural network,the main research contents of this method include the following aspects:(1)For the characteristics of temperature and humidity data derived from the temperature and humidity monitoring system of a pharmacy storage and transportation in Shenyang City,It is processed by the normalization method,and the humidity is interpolated and expanded to improve the accuracy of temperature and humidity prediction.(2)The improved LSTM network consists of an input layer,two hidden layers,and an output layer,and a hidden layer.The LSTM structure in the nested small LSTM cell element replaces traditional iterative prediction and improves training speed.(3)Network training and network prediction adopt adaptive moment estimation(Adam)to adjust network parameters and change the number of network layers to reduce prediction error and improve prediction accuracy.(4)Add a classification module behind the trained network model,classify according to the drug storage temperature and humidity requirements of the new GSP,realize effective monitoring of the medical storage and transportation environment,and provide early warning for pharmacies that do not meet the temperature and humidity requirements.In addition to the above research methods,this paper also compares it with the traditional BP neural network prediction method and the support vector machine model prediction method.The experimental results show that the improved LSTM prediction model is more accurate.