应用马尔可夫链法预测二代粘虫发生程度的探讨

Study on Forecast of Occurrence Degree of Mythimna Separata by Markov Chains Transition Matrix

根据忻府区1983～2006年24年的二代粘虫发生程度的时间序列资料,应用马尔可夫链预测法对该区2007年二代粘虫发生程度进行了预测,结果与大田实际发生情况完全一致。对1994～2006年的历史资料进行回检,符合率为75%,可对二代粘虫的发生程度进行超长期预测。

On the basis of time serial data of occurrence degree of the second generation of Mythimna separata from 1983 to 2006,the Markov chains transition matrix was used to forecast degrees of occurrence of Mythimna separata in 2007,in Xinfu.The forecast result was in accordance with actual occurrence degree.Rate of coincidence was 75% while testing the forecast occurrence degree of Mythimna separate in long term.