An efficient causal structure learning algorithm for linear arbitrarily distributed continuous data

Aim to the linear arbitrarily distributed continuous data, an causal structure learning algorithm BSEM, which is based on simultaneous equations model, was presented. The algorithm merges together simultaneous equations model and local learning. The contribution of this paper is that for linear arbitrarily distributed datasets, BSEM algorithm can effectively learn the causal structure from the datasets. We used the Sociology data to do experiments, and results demonstrated that BSEM displays good accuracy and time performance.