Research and Application of Corrected Weighted Markov Model by Method of Stochastic Optimization Burnishing
Liu Weijie, Jian‐Hua Jia · 2013
In order to overcome the stochastic volatility and state of non-after-effect, we use the b-spline polishing method which have excellent properties of approximation, smoothness, convexity preserving and integrity to randomly optimize and modify weighted Markov model, and eventually corrected weighted Markov model by method of stochastic optimization burnishing. When the model is applied to prediction of annual precipitation, the results show that corrected weighted Markov model by method of stochastic optimization burnishing has higher prediction accuracy, which has important research value and realistic function.