The Application Study Of Bayesian Networks In Digital Forest Eco-station

As a method of describing indeterminate things and indeterminate reasoning, Bayesian Networks is applied extensively in many fields. It is a graphical model for probabilistic relationships among a set of variables and becomes an effective way in the fields of knowledge discovery and decision-making.This thesis analyses both Bayesian Networks and uncertainty and complex relationships between Forest Eco-station's respective factor, then develops a forecast and diagnose system based on Bayesian Networks and database technology, using object-oriented development language Java. It has provided the user a forecast and diagnose platform through the friendly graphical interface and has solved the question which lacks of data analysis, data extract and scientific decision-making in the forest eco-stations at present exists. This thesis takes the tree species's choice as the concrete application and forecasts the planted Pinus tabulaeformis Carr's adaptation in different growth situation. The experiment proves that this system is available and the model has higher precision, It has provided one kind of new method for planting according to the environment and a reference for the tree species's choice, It'll have the instruction significance to the afforestation's success or failure.This forecast and diagnose system based on Bayesian Networks has powerful functions. It not only may be used for forecasting the choose of afforest tree seed, but also may be applied to the forest fire insurance rank forecast, the forest plant disease forecast and so on by establishing different network model according to different research question. This thesis 's Bayesian networks can produce the high confidence level because of its adopting many evidence's evidential propagation method.