Scholarlink

In requirement engineering (RE), correctly classifying functional and non-functional needs has grown to be a crucial effort. The requirements for software are written in natural language text. Machine Learning (ML) -based approaches provide better results than traditional natural language processing. In this study, algorithms for supervised machine learning (SML) that categorise requirements into functional (FR) and non-functional categories are evaluated (NFR). However, comprehensive evaluation of these ML approaches is still required. The machine learning (ML) techniques chosen for this study are Logistic Regression (LR), Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF).The application of ML in RE creates an interesting potential for the creation of new intelligent and knowledgeable systems to assistRE processes and tasks.In this paper, we discuss the treatment of the most distinctive features of FR and NFR, the sampling strategies used in additional data sets, and their impact on classification accuracy. Future work will examine the application of additional algorithms and novel features in an effort to increase the accuracy of the suggested models.

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

Assessing Machine Learning Methods for Segregating Functional User Requirements and Non-Functional User Requirements

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Sonali Rajaram Idate,
  • T. Srinivasa Rao,
  • Dipak Mali,
  • Mrunal Bewoor,
  • Priyanka Paygude

摘要

In requirement engineering (RE), correctly classifying functional and non-functional needs has grown to be a crucial effort. The requirements for software are written in natural language text. Machine Learning (ML) -based approaches provide better results than traditional natural language processing. In this study, algorithms for supervised machine learning (SML) that categorise requirements into functional (FR) and non-functional categories are evaluated (NFR). However, comprehensive evaluation of these ML approaches is still required. The machine learning (ML) techniques chosen for this study are Logistic Regression (LR), Support Vector Machine (SVM), Decision Tree (DT), and Random Forest (RF).The application of ML in RE creates an interesting potential for the creation of new intelligent and knowledgeable systems to assistRE processes and tasks.In this paper, we discuss the treatment of the most distinctive features of FR and NFR, the sampling strategies used in additional data sets, and their impact on classification accuracy. Future work will examine the application of additional algorithms and novel features in an effort to increase the accuracy of the suggested models.