A Retrieval Method of Software Requirements from Japanese Requirements Document with Dependency Analysis and Keywords

Classification of software quality requirements for software quality is a major topic in software engineering. Many methods have been studied using machine learning such as convolutional neural networks and natural language processing. Many studies are being committed to this issue for Japanese language requirement documents as well. However, there is an issue that it is difficult to achieve both precision and recall. This study aims to achieve highly accurate extraction without false positives using keywords and dependency analysis, in light of previous studies.