Study On Prediction Model Of Avian Cholera Based On The Exponential Smoothing Model

Avian cholera is an infectious disease caused by a major violation of Pasteurella multocidafrom chicken, duck, goose, Turkey and other poultry. The disease often presents the septicsymptoms, the incidence and mortality rate is very high, is a serious contagious disease, thenational list of two types of infectious diseases, the disease in most countries in the world havedistribution, at present our country individually in the occurrence and epidemic, harm to ruralchicken Newcastle disease caused by the second, the south the perennially popular, caused greatharm to the industry. In order to early detection and monitoring of avian cholera, this studyestablishes a prediction model and the management and monitoring system.The monthly incidence data based on the official website of veterinary Bulletin released fowlcholera, poultry breeding data and the provincial agricultural statistical yearbook, get choleraincidence.2006January to2011December in Guangdong province fowl cholera incidence andmeteorological data as historical data, the meteorological data as variables, fowl cholera incidenceas the dependent variable, on the basis of variable and dependent variable is calculated by usingthe historical statistical data, then to establish regression equation, then the variable and dependentvariable correlation analysis, the correlation between the two, and calculate the correlationcoefficient, to judge the variable correlation coefficient and correlation variables. In2006Januaryto2010December avian cholera incidence as the original series, the first three observations (i.e.,20061, March2, the monthly incidence rate) of the average as the initial value, generating thesequence plot, according to determine the smoothing coefficient data fluctuation, respectivelyusing an exponential smoothing three model, the two exponential smoothing model, exponentialsmoothing model and double exponential smoothing model, respectively to evaluate severalmodels, comparing the evaluation index and model fitting and forecasting value actual value,finally determine the most suitable model, and the model on the2011January to12monthincidence forecast detection accuracy, prediction model.Results are displayed in the Guangdong Province monthly mean meteorological data, averagetemperature, relative humidity, precipitation, average vapor pressure and fowl cholera incidencewere positively correlated; the maximum wind speed, average pressure had a negative correlationwith avian cholera, which is statistically significant mean air temperature, relative humidity,precipitation, maximum wind speed and fowl cholera incidence of correlation coefficient. Four kinds of exponential smoothing model, an exponential smoothing forecasting model can well fitthe incidence of previous periods of time series, the model of evaluation results for the MSE was0.282×10-10, MAPE0.586, was lower than the other three models, indicate an exponentialsmoothing model can be used to predict Guangdong avian cholera.Avian cholera single exponential smoothing prediction model established in this study andthe effects of meteorological factors analysis provides a new idea and method for the predictionand analysis of the disease, and provide a theoretical basis for the formulation of avian choleraoptimal control strategy, it has certain research and the practical significance.