A tutorial on learning with Bayesian networks

摘要: A Bayesian network is a graphical model that encodes probabilistic relationships among variables of interest. When used in conjunction with statistical techniques, the has several advantages for data analysis. One, because dependencies all variables, it readily handles situations where some entries are missing. Two, can be to learn causal relationships, and hence gain understanding about problem domain predict consequences intervention. Three, both semantics, an ideal representation combining prior knowledge (which often comes form) data. Four, methods networks offer efficient principled approach avoiding overfitting In this paper, we discuss constructing from summarize using improve these models. With regard latter task, describe learning parameters structure network, including techniques incomplete addition, relate Bayesian-network supervised unsupervised learning. We illustrate graphical-modeling real-world case study.