Bayesian Networks have been increasingly used in statistics, decision analysis, artificial intelligence and so on. Because of data mining's strong statistic characteristic, as well as Bayesian Networks originating from Bayesian statistics, it is natural to combine Bayesian Networks with data mining.In this dissertation I dedicate to the research of Bayesian Network's theory and algorithms from a systematic angle. The entire thesis can be divided into four parts.1. Bayesian statistics theory: Expatiate the mathematic principle of Bayesian Networks, Bayesian statistics.2. The reasoning technology of Bayesian Networks: Discuss most reasoning technologies of Bayesian Networks and their computational complexity.3. The learning theory of Bayesian Networks: Discuss the learning principles and processes, as well as the approximate learning methods in the case of incomplete data.4. The C++ pseudocode of reasoning and learning: In this part, I design a kind of data structure for Bayesian Networks' structure and CPTs, construct a Bayesian Networks' C++ class, and write the pseudocode.
|