Informal requirements analysis supporting system for human engineer

As software becomes more complicated and large-scaled, user's demands become various and the expectation level about a software product becomes higher. Therefore, it is very important that a software engineer analyzes user's requirements precisely and applied them effectively in the development step. This paper proposes a requirement analysis supporting system that effectively reduces and revises errors of requirements analysis. The proposed system measures the similarity between requirements specifications or sentences. It extracts the sentences that contain ambiguous words. An indexing method for the similarity measurement combines a sliding window model and a parser model. This method can complement each model's weak points. Using these methods in information retrieval, the proposed system supports a function to trace dependency between documents, improve completeness in a document, reduce inconsistency between sentences, and improve document quality. This paper verifies the efficiency of the proposed system in similarity measurement techniques through experiments, and presents a process for requirements specifications analysis using the system.