APPLICATION OF POISSON PROCESS TO DROUGHT PREDICTION – THE CASE STUDY OF YUCHENG CITY

Abstract. Open-source R language can implement quantitative research using the flexibility, adaptability and simplicity of bayesian inference models. Counts drought which are regarded as realizations “event” in a Poisson process follows an Gamma distribution (prior distribution) case study Yucheng city with support language. That is, annual counts City during 10 years from 1974 to 1983 is process, described by distribution. The (1,1) theta used prior distribution, its posterior calculated be (60,11). sample value obtained random sampling theta, randomly generated for each predicted value. mean results, quantiles value, probability estimation obtain highest density region theta. Studies have shown that model prediction has more provide mean, highest-posterior data, feasible method on basis stochastic thereby providing better reference reduction.