Scholarlink

LLMs have made a big splash on social media by changing how trends catch on and keeping users engaged. LLMs now have the application and understanding of software engineering activities, including requirement engineering, designing, coding, testing and software security. LLMs can help accelerate the process and save cost, time, and enhance security. This paper investigates software engineering activities using LLMs but focuses on requirements evaluation on the generated requirements from different LLMs. The three LLMs used in our work are GPT3.5, through ChatGPT bot; Palm 2, through Poe bot; and Claude 1, through Poe bot. We investigated if LLMs have the understanding and can evaluate the provided requirements according to software quality attributes. We consider the requirements evaluation a main process since all other processes depend on it.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Investigating LLMs Potential in Software Requirements Evaluation

  • 章节

  • 发表日期：

    2025年

查看全文
  • Najlaa Alsaedi,
  • Ahlam Alsaedi,
  • Amjad Almaghathawi,
  • Mai Alshanqiti,
  • Abdul Ahad Siddiqi

摘要

LLMs have made a big splash on social media by changing how trends catch on and keeping users engaged. LLMs now have the application and understanding of software engineering activities, including requirement engineering, designing, coding, testing and software security. LLMs can help accelerate the process and save cost, time, and enhance security. This paper investigates software engineering activities using LLMs but focuses on requirements evaluation on the generated requirements from different LLMs. The three LLMs used in our work are GPT3.5, through ChatGPT bot; Palm 2, through Poe bot; and Claude 1, through Poe bot. We investigated if LLMs have the understanding and can evaluate the provided requirements according to software quality attributes. We consider the requirements evaluation a main process since all other processes depend on it.