Research And Implementation Of Prediction Model For Class B Infectious Diseases Based On Machine Learning

As the development trend of infectious diseases is still unclear,this means that it is not easy to predict the incidence of infectious diseases.With the deepening of machine learning and deep learning research,especially the extensive research of LSTM cyclic neural network,many unpredictable problems have been better solved.This thesis finds that the LSTM cyclic neural network has not been applied in the field of infectious disease prediction.This article is a bold attempt to apply LSTM neural network to the field of infectious disease prediction.Through the use of weather data,economic data,population data,etc.,and combining these data with infectious disease cases data to conduct research on infectious disease prediction,a good predictive effect of infectious diseases has been achieved.Firstly,the existing data is analyzed through data preprocessing,and useful characteristic factors are extracted.For example,the data analysis shows that the incidence of infectious diseases is not only the above,but also age factors,gender factors,occupation factors and so on.In this thesis,through the comparison experiment between ARIMA and LSTM,it is found that the LSTM model is superior to the ARIMA model when predicting the incidence of all infectious diseases,and all data is divided into predictions of multiple time periods,for example,forecast by month,week and day.In the selection of the forecast time period,this thesis finds that the effect of the weekly forecast is optimal.In predicting the incidence of infectious diseases in the next week,the RMSE index of the LSTM model is 188.59,while the RMSE index of the ARIMA model is 336.88.Based on the premise that the effect of weekly prediction is optimal,this study extracts the dataset of a single disease to make predictions.In the case of consistent data,predicting the incidence of hepatitis,tuberculosis,syphilis and other ARIMA.The pros and cons of the model and the LSTM model.From the perspective of prediction results,the LSTM model is superior to the ARIMA model.The above experimental results show that the LSTM circulating neural network is more accurate and more suitable for the prediction of Class B infectious diseases.After the above comparative experiments,this thesis decided to use LSTM circulating neural network to predict the incidence of infectious diseases,and made a variety of indicators of the prediction results into a complete,high-performance visualization system.We believe that this study can eliminate problems such as reporting delays and inaccurate predictions in existing infectious disease surveillance systems,thereby minimizing social costs.The system has also been tested within the Sichuan Provincial Center for Disease Control and Prevention.