Research On The Prediction Of Expenditure And Hospitalization Numbers Of Hospitals Based On Improved GM(1,1) And SVR Model

With the rapid development of information technology and continuous improvement of hospital facilities construction,the use of information technology to manage,analyze and predict hospital data and provide reference for the development of hospitals has received more and more attention from the society.In recent decades,large-scale hospitals in our country have experienced rapid development.There has been a significant increase both in hospital spending and in hospitalization.People want to be able to make more accurate predictions of the expenses of large hospitals so that they can reduce their expenditures and increase efficiency.People also hope to make accurate predictions about the number of hospital admissions in large hospitals,so as to provide reference for the hospital's workload arrangements,and the storage and preparation of medicines and consumables.From ancient times to now,the prediction of uncertain events is always a great human desire.With the progress of statistics and computer science in modern times,people have a more scientific method of predicting the development of things.By analyzing the laws of the data itself and the impact between the data,their changing trends are obtained.Mathematical methods are used to establish a function that describes the law of data change and the degree of influence between the data,and then the prediction of the future is made through the derivation of the function.Therefore,the accuracy of the prediction mainly lies in the accuracy of the function's description of the data pattern.The higher the accuracy,the more accurate the prediction.Hospital expenditure grows steadily,and can be predicted based on the regularity of the data itself.The number of hospital admissions by month is random,so it is necessary to consider various factors,analyze the factors affecting the number of admissions,and establish mathematical models before making predictions.This article will use two kinds of predictive models to analyze and solve the two problems respectively.In view of the existing deficiencies in the existing research,this article makes predictions for hospital expenditure and hospital admissions based on two prediction models,and improves and optimizes the models.The specific work and contributions of this article are summarized as follows:1.A method for large hospital expenditure forecasting based on adaptive growth rate model GM(1,1)(AG-GM(1,1))is proposed.In this paper,two models have been established through in-depth study of large hospital expenditure data series and GM(1,1)model.The background value calculation formula for the prediction model suitable for the high-growth sequence is modified with a complex Simpson integral formula to fundamentally improve its accuracy,and then the initial value selection method is changed,and the two optimizations are combined to improve the model.The classical model background value constructor of the prediction model applied to the general growth sequence is weighted and optimized,and the traversal method is used to find the optimal solution,then the initial value selection method is combined to form a model.An algorithm was designed to calculate the growth rate of hospital expenditure data and use its value to determine which model will be chosen for prediction.Then balance factors are calculated for large fluctuations and their prediction results are corrected using the balance factor.The effectiveness of the proposed method is proved by experiments,and the accuracy is improved in the prediction.2.A method for predicting the number of inpatients in large hospitals based on a weight-restricted Drosophila-optimized Support Vector Machine Regression(WF-SVR)model was proposed.Support vector machine regression(SVR)model has been optimized and improved:a fruit fly optimization algorithm is used to optimize the parameters,so that the parameter selection process is more high-quality and efficient,and the time consumption is reduced while improving the quality of the parameters.The weights of punishments are weighted to make them have different punishments for the samples that appear at different times during training,which more accurately reflects the influence of time-series samples on future data.This paper analyzes the influencing factors of the number of hospital admissions,determines six main influencing factors,and uses an optimized support vector machine regression model to predict the number of hospital admissions.This article uses datasets made up of data from some hospitals and their cities to conduct experiments and compare experimental results with basic methods.The experimental results show that the optimized prediction model has better prediction ability.