链图模型的结构学习研究-应用数学专业论文.docxVIP

山东 山东师范大学硕士学位论文 万方数据 万方数据 立性，和学习复形的算法，对整体骨架定复形，得到潜在链图的模型。 关键词: 链图; 条件独立; 边缘独立性; 递归算法. 分类号: O157.5 2 Structural Learning of Chain Graphs Model Luo Panpan Shandong Normal University, Jinan, Shandong, 250014 People’s Republic of China ABSTRACT Causal Diagram Model is the main tool for inference of causal. It has the stem part named simple causal diagram model which is been seen as a useful discipline be- cause of its popularity and practicability. It is only in recent years that scholars begun the specific research into Causal Diagram Model. But until now, the system is still not perfect.Causal Diagram Model includes the structures of Markoff and Bayes. For the structure of Markoff, it is used for the qualitative and quantitative. The structure of Markoff is a network structure with no phase, where the nodes represent variables and the sides represent correlation of two variables. Here, we use the structure with no phase to discuss the independence of two variables which is also been used on recognition of Speech, transformation of Text-to-Speech and Part-of-speech tagging. For the structure of Bias, it is based on the theorem model of probability inference which is a process of obtaining the other informations of statistics of model from the known variables. The structure of Bayes is a phase diagram without rings, where the sides with directions rep- resents causation and the nodes still represent variables. Thus, Bayesiannetwork is come up from probability inference to solve the uncertainly and incompletely of problems. It has obvious advantage to solve the breakdown of equipment which was caused by the uncertainly and relevance of complex equipment. Also, it is used widely in other areas, such as social science, economics and computational science. In this article, I mainly talked about how to study the local structure of a Chain Graphs with its non-observed quantities were marginalized and how to recover the whole network graph from two local structures. Also, I ex