Bayesian Network Structure Learning and Application

With the continuous development of artificial intelligence technology and information technology, a large number background data are constantly generated. How to obtain effective useful in complex group becomes important meaningful. The traditional Bayesian network can represent probability distribution variables from based on graphical models. It has relatively clear reliable reasoning ability decision-making mechanism. However, structure serious shortcomings recognition accuracy corresponding key data, so efficiency algorithm is seriously low. Based this, this study adds an adaptive genetic with causality original structure, as optimize strategy its operation, quantitatively describe order nodes, creatively arrange nodes by using node priority, initialize initial architecture this. Finally, initialized through exchange score correction, get final learning results. In study, convolution neural database verified experiment. experimental results show that given proposed about 10% higher than accuracy, basically cover algorithms, hypotheses, verification network, level; obvious advantages bibliometrics.