A method for text information acquisition based on Bayesian network

It is proposed an algorithm, which is on the basis of VGH location method, for text location in complex settings based on Bayesian networks to deal with the uncertainty problems. In this article, we use Bayesian networks to set the statistical relations between features and text objects and analyze probability of text existence. We design three kinds of Bayesian networks to identify regularities of convex peak or concave valley and text existence, and obtain the condition probability tables through condition expectation learning algorithm. The experiment results show efficiency of algorithm.