Causal Discovery Based On Genetic Algorithm

The causal relationship is the main relationship between things, the effective discovery of the causal relationship contributes to guide our daily work and decision-making.But in real life,it is difficult to determinant the causal relationships between things.In recent years, with the rapid development of information technology and the extensive using of information system, enterprises or units have accumulated a large number of data during operations, so finding the causal relationship has become a popular research direction.In this thesis, according to the problem of the causal discovery of data, firstly it would be introduce the Bayesian network to express and describe the causal relationship,and turns causal discovery problem into learning problems of the Bayesian network structure;then we design and program the classic K2algorithm by using genetic algorithm,it can be improve the operating efficiency during the original algorithm.Finally, we choose a case from a china mobile company based on actual operational data, it using the Bayesian network which is based on genetic algorithms to analyze the reasons of the off-grid probability of the corporate clients,it using the Bayesian network which is based on genetic algorithms to analyze the reasons of the off-grid probability of the corporate clients,the results will show the algorithm has a certain validity and applicability.