Abstract
It is important to examine the risk without omission at the project early stage. If only enumerating the risk and examining it can consider the relation of the risk, efficient risk management can be done. In this research, The risk is extracted by using the text mining and Bayesian inference and the causal relations between risks are presumed, and it proposes the technique for making it to visible on the network, and the verification that uses the failure case with the management crisis data base of the venture company is done.