Retrieving Information from a Document Repository for Constructing Assurance Cases

We address the problem of constructing an assurance case by presenting an approach to extract information from a large set of documents. In the proposed approach, document retrieval and formal concept analysis techniques are systematically combined for assisting users to explore relevant information from huge data set and to understand a number of concepts in such data set with the relation among them. We perform an experiment with a data set from an open-source software development project, in order to evaluate the effectiveness of our approach. The experimental results suggest that the proposed approach can be effective in terms of reducing the time and the cost for constructing assurance cases with acceptable confidence level, indicated by some assurance case quality metrics.