A SMILE web-based interface for learning the causal structure and performing a diagnosis of a Bayesian network

Bayesian networks (BNs) are probabilistic graphical models that are widely used for building diagnosis- and decision-support expert systems. The construction of BNs with the help of human experts is a difficult and time consuming task, which is prone to errors and omissions especially when the problems are very complicated. Learning the structure of a Bayesian network model and causal relations from a dataset or database is important for large BNs analysis. This paper focuses on using a SMILE Web-based interface for building the structure of BN models from a dataset by using different structural learning algorithms. In addition to building the structure of BN models, a SMILE Web-based interface also provides the feature set of Bayesian diagnosis for the user. The Web application uses a novel user-friendly interface which intertwines the steps in the data analysis with brief support instructions to the Bayesian approach adopted. A SMILE Web-based interface has been developed based on SMILE (Structural Modeling, Interface, and Learning Engine), SMILEarn, and SMILE.NET wrapper.