Eliciting Relations from Natural Language Requirements Documents Based on Linguistic and Statistical Analysis

Abstract

Requirements are usually presented as Natural Language based documents. In the conceptual modeling phase, requirements are collected from different stakeholders and analyzed by requirement engineers. However, the size of the requirements documents can become very large, and the modeling process is quite time consuming and resource consuming. In order to solve this problem, much has been written on the processing of requirements documents to yield conceptual models. In this paper, we proposed an approach for identifying and extracting relations in a range of requirements documents with three steps: text analysis, entity extraction and relation mapping. If the entities in the relation are quite close to each other, for example, in the strategic dependency relationship, we will define a set of linguistic patterns used for identifying relations and propose a matching algorithm of semantic automata to extract the relation. Based on this approach, we developed a system to automatically generate the strategic dependency model of i framework and the activity model from Chinese requirements documents. A series of experiments were conducted to evaluate the performance of the automated requirements analysis system. The results show that the system achieves high recall with a consistent improvement in precision, which demonstrates the applicability of our approach.