Abstract
In late years a thing in the present conditions that the success rate of the software development project in the company of our country is extremely low is a well-known fact. I think that the risk identification at the estimate stage is effective to plan the improvement of this success rate. However, the identification precision is low to distinguish the risk at this stage because information about the project is few to quality / quantity. Therefore I suggest the risk identification technique that utilized Collaborative Filtering to enable highly precise recommendation work in the information less to support information less by this report.