Application of time series analysis in the prediction of rice blast
Mei Zhang · Journal of Northwest A&F University · 2011
【Objective】 The prediction effect of rice blast was studied to provide reference to control rice blast by time series analysis model.【Method】 Time series analysis model was established through moving average,exponential average and period extrapolated method by variance analysis,according to the incidence of plant areas of rice and apperance of rice blast in Jiange district Sichuan province from 1980 to 2007.The forecast result was analysed and used in the appearance predication of rice blast in 2008.【Result】 It was shown that the time series analysis linear model including moving average,exponential average and period extrapolated method by variance analysis can preferably fit the trend of rice blast.Comparing the measured value with the estimated value,their accuracy of testing-forecast from 1980 to 2007 was 97.7%,96.1% and 99.8%,respectively;relative errors in 2008 were 14.5%,18.1% and 8.4%,respectively.【Conclusion】 Time series analysis model can be applied in a long-term forecasting for rice blast.