Scholarlink

<p>Software testing is a critical phase in ensuring the reliability of applications, where the generation and prioritization of test cases are paramount. Traditional approaches often lack the ability to dynamically adapt to user-driven requirements or utilize real-world feedback effectively. This paper proposes an innovative hybrid framework that leverages Long Short-Term Memory (LSTM) network for generating test cases and K-Nearest Neighbors (KNN) clustering for prioritization. Customer feedback is incorporated as a key factor, ensuring alignment with user expectations and criticality. LSTM is capable of learning temporal dependencies in sequences, which is ideal for textual data. KNN Clustering works well in case of high-dimensional feature vectors, such as the ones extracted by an LSTM, and want to group similar data together based on distance metric. Experimental results demonstrate the framework’s superiority in enhancing fault detection rates, reducing execution time, and improving customer satisfaction as compared to the existing studies.</p>

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

LSTM-KNN: a novel and hybrid prediction framework to enhance prioritization of regression test cases

  • 期刊论文

  • 发表日期：

    2025年8月29日

查看全文
  • Umakanta Dash,
  • Arup Abhinna Acharya,
  • Satya Ranjan Dash

摘要

Software testing is a critical phase in ensuring the reliability of applications, where the generation and prioritization of test cases are paramount. Traditional approaches often lack the ability to dynamically adapt to user-driven requirements or utilize real-world feedback effectively. This paper proposes an innovative hybrid framework that leverages Long Short-Term Memory (LSTM) network for generating test cases and K-Nearest Neighbors (KNN) clustering for prioritization. Customer feedback is incorporated as a key factor, ensuring alignment with user expectations and criticality. LSTM is capable of learning temporal dependencies in sequences, which is ideal for textual data. KNN Clustering works well in case of high-dimensional feature vectors, such as the ones extracted by an LSTM, and want to group similar data together based on distance metric. Experimental results demonstrate the framework’s superiority in enhancing fault detection rates, reducing execution time, and improving customer satisfaction as compared to the existing studies.