Predict The Incidence Of Hepatitis B Based On Grey And Neural Network Model

Along with the developing of the economy and the great advantage by science, people pay more and more attention to predictions in medical treatment and public health area. Nowadays economic effectiveness is a question which is worth more consideration, how to enhance economic effectiveness in medical area, prediction made a great effort. First of all, in the aspect of infectious disease, predict the incidence of infectious diseases can guide the medical and health institutions distribute the human and material resources in a timely manner to control infectious disease outbreaks and reduce the harm. Second, though the predict outpatient services hospitalization rates of a the hospital and other data reflecting the extended capabilities of a hospital as well as the level of medical technology medical resources can be distribute rationally, evaluation the level of the hospital’s business more correct.We used to use regression methods, exponential smoothing, ARIMA forecasting method, the gray prediction method, neural network prediction method on the prediction of infectious diseases. These prediction methods have their own advantages, but there are also insufficient. How to improving forecast accuracy, reducing error is the main problem encountered in the forecast.The gray model system use the date generated by the cumulative to modeling, it is in some way weaken random of the raw data and make us find the variation of the data easily, the sample size required by the modeling is small, and short-term forecasts have high precision. But at the same time it is difficult to deal with historical data when it is anomalies, and use grey model system to modeling dependence on historical data. Predict will turn to a large deviation when the data appear a larger change.The neural network has the ability to imitate a variety of functions, can pressing any continuous nonlinear function in any closed interval, can avoid information distortion generated by the other system data identification method, with good adaptation and self-learning ability. But the need of sample size is large, the process of learning convergence is slow, and very sensitive to initial weights, can easily converge to local minimum value, the network learning and memory is instability and other issues.Use gray neural network model to predict the time series compared with the gray model: accuracy is higher, and the error is controlled. Compared with the neural network model, the calculation amount is small can also achieve higher prediction accuracy in the case of fewer sample. This will not only make full use of gray models modeling’s advantage such as the required sample data is small, the principle is simple, computing is convenient, short-term forecast more accuracy but also the advantages of a neural network such as parallel computing, fault-tolerant and adaptive ability, can reflect the data volatility change. This paper uses the combination of gray neural network method to predict the incidence of infectious diseases, and compared with gray model and neural model. The result show that gray neural network method has a better predicts effect.