Research On The Examination Path Planning For Patients In Hospital Based On The Combination Of Artificial Neural Network And ARIMA Model

[Purpose] Taking the outpatient department as the starting point,for the patients who need to complete the examination and inspection items at the same time,we aims to plan a route to help patients complete all projects within the shortest time comprehensively considering the time to walk,complete the examination and queue in the hospital.[Methods] The ARIMA model and Artificial Neural Network were used to calculate the patient’s traffic speed in the hospital and the queuing time of the department.At the same time,combined with the average time required to complete each examination project,the execution department with short time was selected in turn,until all the projects that need to be implemented are all finished,and finally we summed up into a path that takes the shortest time to recommend to the patients.[Results] The ARIMA model was used to predict the outpatient volume of dermatology and the queuing time of laboratory departments.The MAPE,RMSE and MAE of the two models were 8.49,270.03,203.83 and 1.30,16.68,10.02,respectively.The prediction model of queuing time of ultrasound department by Artificial Neural Network reached 95% accuracy,and its MAPE,RMSE and MAE were 3.33,5.88 and 3.95 respectively.In addition,the historical average time consumption of patients under different specific conditions is compared with the time consumption after model optimization,and it is found that the planned paths save patients’ time in varying degrees.[Discussions](1)The number of dermatology outpatients in the hospital is so large that makes the gallery crowded and the patients’ passing speed is limited.In addition,the queuing time of patients in ultrasound imaging department has an obvious variation tendency,which can be alleviated by effective guidance and peak shunting;the queuing situation of the medical laboratory in the morning is more serious,which can be improved by increasing the number of open windows.(2)Compared with other similar models,the prediction model of this research has certain advantages,and its estimated time is shorter than the average time consumed by patients,thus it has a certain practical value.(3)Potential environmental factors such as air quality can be included in the study to further improve the accuracy of prediction.