Managing User Requirements Engineering with Automated Speech Recognition and Natural Language Processing Tools

The automation of software development processes has advanced significantly in the 21st century, yet software documentation practices remain largely traditional. End users often face challenges in eliciting and classifying requirements, which are critical for effective software development. This study aims to address these challenges by developing an automated tool that facilitates user requirements elicitation and classification. We overcome the problem using speech recognition technologies and natural language processing (NLP) algorithms to streamline these processes. The tool was designed and developed in two phases: the logical design phase and the physical implementation phase. A webbased approach was utilized to generate statistical classification results, enabling the categorization of user requirements into functional and non-functional classes using predefined tags. Functionality and accuracy testing were conducted to evaluate the tool's performance. The results demonstrated the tool's effectiveness, achieving 80 % accuracy in two-class classification and met the study's objectives. These findings suggest that the proposed automated tool provides a reliable solution for supporting software engineers in managing user requirements elicitation and classification, enhancing the efficiency and accuracy of the requirements engineering process.