An Incremental Approach for Sparse Bayesian Network Structure Learning

A Bayesian network is a graphical model which analyzes probabilistic relationships among variables of interest. It has become a more and more popular and effective model for representing and inferring some process with uncertain information. Especially when it comes to the failure of uncertainty and correlation of complex equipment, and when the data is big. In this paper, we present an incremental approach for sparse Bayesian network structure learning. In order to analysis the correlation of heating load multidimensional feature factor, we use Bayesian network to establish the relationship between operating parameters of the heating units. Our approach builds upon previous research in sparse structure Gaussian Bayesian network, and because our project requires us to deal with a large amount of data with continuous parameters, we apply an incremental method on this model. Experimental results show that our approach is the efficient, effective, and accurate. The approach we propose can both deal with discrete parameters and continuous parameters, and has great application prospect in the big data field.