Objective To describe the procedure of building Elman neural network model, and explore the value of potential application of the above model. Methods Monthly incidence of hemorrhagic fever with renal syndrome(HFRS) in China from 2004 to 2013 was used to build Elman neural network model and SARIMA model and forecasted the monthly incidence of HFRS in China from January 2014 to September 2014. The fitting and prediction effects of the two models were compared. Results For training sample, MAE, MAPE and RMSE of Elman neural network were 0.0088, 0.1191 and 0.0127 respectively; MAE, MAPE and RMSE of SARIMA model were 0.0111, 0.1268 and 0.0206 respectively. For predicting sample, MAE, RMSE and MAPE of Elman neural network were 0.0079, 0.1180 and 0.0096 respectively; MAE, RMSE and MAPE of SARIMA model were 0.0178, 0.2778 and 0.1861 respectively. Conclusion Elman neural network fits and forecasts the HFRS incidence trend in China well, and the fitting and prediction effect is superior to the SARIMA model, which is of great application value for the prevention and control of hemorrhagic fever with renal syndrome.

Objective To analyze the epidemiological features of hemorrhagic fever with renal syndrome (HFRS) and associated environmental risk factors for HFRS and associated environmental risk factors for HFRS in Liaoning province, China during 2005-2007, and to provide a scientific basis for HFRS control measures. Methods The epidemic data of HFRS in Liaoning province were collected. Analysis was performed to determine the correlation between the epidemic features of HFRS and environmental factors such as mean temperature, relative humidity, rainfall, sunshine, urban rodent density, rural rodent density, and virus?carrying rate. Results There were 7298 cases of HFRS in Liaoning province from 2005 to 2007, and 78 of them died. The mean annual incidence of HFRS was 5.78/100 000, and the mortality was 0.06/100 000; the incidence and mortality were higher in males than in females; 59.55% of the cases and 69.23% of fatal cases were aged 35-60 years; 61.98% of the cases and 56.41% of fatal cases were farmers. The peak of incidence appeared mainly in November to January and March to May, while the trough period was in July to October, showing the seasonal characteristics in mixed epidemic area; the mean annual incidence of HFRS was relatively high in the cities of Benxi (13.70/100 000), Huludao (12.92/100 000), Jinzhou (11.30/100 000), Dandong (10.21/100 000), and Fushun (9.84/100 000). The incidence of HFRS was negatively correlated with temperature but positively correlated with rainfall, rural rodent density, and virus-carrying rate; the Spearman rank correlation coefficients were -0.351, 0.400, 0.449, and 0.377, respectively, and the P values were 0.023, 0.009, 0.003, and 0.016, respectively. Conclusion In Liaoning province, HFRS is prevalent mainly in winter and spring and among young male farmers. The prevalence of HFRS is closely related to temperature, rainfall, rural rodent density, and virus-carrying rate in the same year.

【Abstract】 Objective To study the relationships of meteorological factors, animal host and hemorrhagic fever with renal syndrome （HFRS） incidence, and construct mathematical model for the forecast of HFRS. Methods Firstly, air pressure, air temperature, relative humidity, precipitation, sunshine duration and sunshine percentage were selected from all meteorological factors of Huludao city. Secondly, Pearson, Kendall and Spearman correlation analyses were used to describe the relationships among meteorological factors, animal host situation including rodent density and viral carriage of rodents and HFRS incidence. Thirdly, Bayesian discrimination analysis (BDA) was adopted to forecast HFRS incidence on the premise of meteorological factors and animal host formation as explanatory variables. Results There was the close relation between rodent density and annual HRFS incidence(r＝0.738, P＝0.000), and the rodent density was also influenced by sunshine duration, sunshine percentage and precipitation.  A positive correlation was found between rodent density and sunshine time(r＝0.494, P＝0.016), and the correlation between rodent density and precipitation was negative(r＝-0.350, P＝0.101). The step wise BDA and all variables discrimination analysis had all good effect on the forecasting of HFRS based on meteorological factors and animal host data. The accuracy rate of fitting and leave?one?out (LOO) cross-validation of stepwise BDA all reached 82.6%(19/23) , however, that of  fitting of all variables BDA was 90.9%(20/22) and 81.8%(18/22) for LOO cross-validation. For next year incidence prediction, the accuracy rates of fitting and LOO cross-validation step-wise were all 86.4%(19/22) for step-wise BDA, while for all variables BDA, its accuracy rate of fitting was 100%(21/21) and that of LOO cross-validation was 57.1%(12/21). Conclusion HFRS incidence was related to animal epidemic situation which was influenced by meteorological factors. Stepwise BDA offered useful information in the discrimination and forecasting of HFRS incidence, which had a good application in the future.

Objective To study the superiority and application prospect of support vector machine(SVM) on the forecast of the incidence of hemorrhagic fever with renal syndrome(HFRS).Methods Firstly,the routine meteorological data of Huludao city including average air pressure,average temperature,relative humidity,precipitation and sunshine time and the epidemiologic information of animal disease including rodent density and rodents borne virus from 1984 to 2006 were used as predictable variables.All the variables were limited to the range from 0 to 1.The whole data atlas were separated into training atlas and test atlas.The test atlas were made up of 1/3 individuals(trunc) randomly sampled from data atlas,and other samples were composed of training atlas.Secondly,SVM was applied to the HFRS incidence prediction and the SVM model was constructed by software R2.60.Finally,the performance of SVM,back-propagation(BP) and radial basis function(RBF) Neural Networks were compared by computing the sum square error(SSE).The above procedures were repeated for 10 replications.Results The mean and standard diviation of SSE of SVM for training atlas was(0.031±0.009),while those of BP and RBF neural network were(0.074±0.030) and(0.082±0.018),respectively.For the test atlas,the mean and standard diviation of SSE of SVM was(0.067±0.021),while those of BP and RBF neural network were(0.073±0.022) and(0.089±0.036),respectively.Conclusion As a new pattern recognition method developed on the basis of statistics theory in recent years,SVM had higher forecast precision and stronger generalization ability to solve the small sample size and the indentification of nonlinear and high-dimension model,SVM was reliable for the prediction of HFRS incidence,which could serve as a reference method for the HFRS prediction.

Objective To study the superiority and application of generalized regression neural network(GRNN) combination forecast model in the forecast of hemorrhagic fever with renal syndrome(HFRS) incidence. Methods Establish the GM(1,1) model and auto regressive integrated moving average(ARIMA) model based on the data of HFRS of Dandong, Shenyang and Chaoyang Liaoning province, from 1990 to 2001 respectively. The forecasting values of the two models were used as input of GRNN. Train the sample and forecast the value. Compare the forecasting effect of the three models. Results The mean error rate(MER) of GM(1,1) model, ARIMA model and GRNN combination model for Liaoning province were 13.5143%, 25.0814% and 5.5755% respectively. The R ² of the three models were 0.8961, 0.6997 and 0.9837 respectively. The MER of GM(1,1) model, ARIMA model and GRNN combination model for Dandong were 19.7329%, 20.6275% and 14.0789% respectively. The R ² values of three models were 0.8112, 0.7628 and 0.8750 respectively. The MER of GM(1,1) model, ARIMA model and GRNN combination model for Shenyang were 15.1421%, 18.0584% and 14.3592% respectively. The R ² values of three models were 0.8757, 0.7889 and 0.8585 respectively. The MER of GM(1,1) model, ARIMA model and GRNN combination model for Chaoyang were 51.5090%, 28.6593% and 28.5927% respectively. The R ²values of three models were 0.7863, 0.8291 and 0.7753 respectively. The forecasting efficacy of combination model for Liaoning province was better than other two single models. For the forecasting efficacy of Shenyang, the GRNN combination model and the GM(1,1) model were similar, and the ARIMA model was the worst. The incidence of HFRS for Chaoyang is not fit for the establishment of the models we mentioned above. Conclusion GRNN combination model had more advantage in the forecast of small sample and the forecasting efficacy was better than GM(1,1) model and ARIMA model, which had practical value in the treatment of time series data such as the incidence of HFRS.

Objective To study the superiority and application prospect of generalized regression neural network(GRNN) which is used in forecasting the incidence of hemorrhagic fever with renal syndrome(HFRS). Methods Use meteorological data, including average temperature, relative humidity, precipitation and sunshine time, and epidemiologic information of animal diseases, including rodent density and viral carriage of rodents from 1984 to 2002 as the input of neural network. Use the incidence of HFRS from 1985 to 2003 as the output of neural network. Construct the GRNN forecasting model and BP neural network forecasting model respectively with the neural network toolbox of Matlab7.0. Fit and forecast the sample and compare the performance between the two different neural networks. Results The optimize smooth factor of GRNN is 0.35; the hidden layers of BP neural network is 6. From the fitting effect, the MER of GRNN and BP neural network are 25.42% and 25.55% respectively; their r ² are 0.9438 and 0.9729. On the whole, the fitting effect is satisfactory, and the difference of the two neural networks is not very significant. From the forecasting effect, the MER between the two neural networks are 4.90% and 15.16% respectively. The MER of GRNN is less than the MER of BP neural network; their r ² are 0.9897 and 0.9516. Conclusion GRNN is more superior in small sample forecasting than BP neural network, and the forecasting effect is better. GRNN has practical value in solving epidemic problem which has complicated influencing factor such as HFRS.

Objective To study the application of back propagation (BP) artificial neural network model in prediction for incidence of hemorrhagic fever with renal syndrome(HFRS).Methods Meteorological data,including average temperature,relative humidity,precipitation and sunshine time,obtained from Shenyang Municipal Meteorological Bureau,and epidemiologic information of animal diseases,including rat density and viral carriage of rats obtained from Shenyang Municipal Center for Disease Control and Prevention,were collected as input of artificial neural networks.And,incidence data of HFRS in Shenyang during 1984 to 2003 were collected as output of artificial neural networks.A predictive model of BP artificial neural networks was established using the data during 1984 to 2001 with STATISTICA Neural Network(ST NN) software.The