Medical Cold Chain Data Analysis And Research Based On Deep Learning

As for solve the low-temperature drugs in the storage,transportation and other medical cold chain links,to ensure the efficacy of drugs,and because of the drug storage time is relatively short shelf life of medicines and vulnerable to the surrounding environment,Therefore,this thesis designs a pharmaceutical cold chain data monitoring system based on deep learning.The system collects a variety of basic information in the warehouse and uses convolutional neural network and other algorithms,the server analyzes whether the temperature and humidity data recorded by the temperature and humidity recorder in the warehouse have the same-direction change,invalid,excessive temperature,excessive humidity,reverse change,sensor failure,Frost,at the same time exceeding the standard,man-made changes and other ten cases,which reflects the storage of drugs that solve people's medication safety problems also help the relevant departments to implement effective supervision.In this thesis,the convolution neural network algorithm is studied in depth,by constantly changing the number of hidden layers in the convolutional neural network algorithm and the size of the convolution kernel and decimation ratio,the most suitable network parameters of the system are obtained,So that the system uses convolution neural network algorithm to quickly and efficiently classify the temperature and humidity data,extract the effective features,diagnose different fault conditions,and make it effectively applied to the drug regulatory system.The input data selected in this thesis is 2 × 288,and 5 convolutional layers,4 descending layers and 2 fully connected layers are designed,and the parameters of each layer are adjusted accordingly.In order to make it easier for authorities to check each fault condition,this thesis also systematically designs a phone APP of the cold chain temperature and humidity detection system.This APP is a fast-browsing software running on a smart phone,designed primarily for setting up a development environment based on the Android mobile platform.The phone APP for the relevant departments to query the temperature and humidity data correlation analysis results for the regulatory work to provide real,timely and effective data.The experimental results show that deep learning represented by convolutional neural network algorithm is feasible to diagnose the fault of medical cold chain data,and the outstanding characteristic of this algorithm is the high detection accuracy.