Test Case Prioritization Through Clustering: A Data-Driven Approach

Abstract

In software development, maintaining software quality and reliability through thorough testing is vital. However, as software systems become increasingly complex, managing large volumes of test cases presents significant challenges [1]. To tackle this problem, effective test case prioritization is necessary to ensure that the most critical tests are executed first    [2]. This paper introduces an innovative approach to test case prioritization by combining clustering techniques, specifically K-means, with machine learning algorithms. We explore how K-means clustering can group similar test cases to improve prioritization efficiency [3][4]. Furthermore, we examine the performance of several machine learning models, including Decision Trees (DT), Random Forests (RF), and Neural Networks (NN), comparing their results against traditional methods. The study evaluates these approaches using diverse datasets and metrics such as the number of executed test cases, fault detection rate, and execution time [5]. Experimental findings demonstrate that integrating K-means clustering with machine learning techniques can enhance prioritization by reducing test execution efforts while preserving or even boosting fault detection capabilities. We also highlight the limitations of the proposed method and suggest future research opportunities aimed at further optimizing test case prioritization through advanced machine learning strategies. Overall, this framework offers important contributions toward developing more effective and reliable software testing processes.