Research On Prediction Model Of Hospital Outpatient Volume

The hospital outpatient department provides patient diagnosis,patient treatment and medical care.It is an important component of the hospital and the department where the patient's visit to the clinic first arrives,reflecting the overall image of the hospital.Outpatient statistics are important statistical indicators of hospital work.Accurately predicting the number of outpatients can provide decision-making basis for hospital administrators,and is the basis for better utilization of resources and improvement of outpatient treatment.Therefore,this paper first analyzes the outpatient volume by using several commonly used single models.Finally,based on these analyses,a combined forecasting model is established.The error analysis is used to obtain the statistical model with the highest accuracy,and the trend of hospital outpatient volume is scientifically and accurately analyzed.Providing decision support for hospital administrators and rational allocation of medical and health resources,thereby further improving the efficiency of hospital outpatient services,saving manpower and material resources,and facilitating outpatient services in hospitals.This study first descriptively analyzed the monthly outpatient statistics of a top three hospital in Kunming from 2000 to 2016,and then used a single long-and short-term memory neural network and traditional ARIMA model to model and train appropriate predictive models and conduct Forecast,two sets of prediction results are obtained respectively.Finally,combined predictions were performed using the EMD-LSTM and EMD-ARIMA combination models,and four different prediction results were obtained.Subsequently,several modeling methods were compared and analyzed.The model was evaluated by Mean Absolute Error(MAE),Root Mean Squared Error(RMSE),Mean Absolute Percent Error(MAPE)and R-square.Finally,by analyzing the prediction results,it is concluded that,firstly,Empirical Mode Decomposition(EMD)can improve the accuracy of the prediction model to a certain extent,but the degree of improvement is limited.Second,for the outpatient volume time series data,the combined model The prediction effect is better than the single model that makes up the combined model.Finally,this paper analyzes the shortcomings of the research and looks forward to the research on improving the time series prediction analysis of outpatients.