Scholarlink

Software defects are well known in software development and can create many issues for developers and users. Consequently, researchers have used distinctive methods to alleviate the implications of these flaws in source codes. Software defect prediction (SDP) provides observable results to development groups when contributing to industry decisions that predict defect code areas to support developers in identifying errors and arranging their testing activities. Thus, machine learning (ML)-based technique is designed for SDP. First, input software data are obtained and forwarded to the feature selection process. Feature selection is performed utilizing Information Gain (IG). The system forwards the selected features to the SDP process, where it applies the rider optimization algorithm-based neural network (RideNN), trained using the designed Fractional Pufferfish Optimization Algorithm (FPOA), to carry out SDP. The established FPOA was devised by incorporating Fractional Calculus (FC) and the Pufferfish Optimization Algorithm (POA). Finally, SDP analysis of SDP is performed with the learning data. Moreover, the FPOA-RideNN was evaluated using several metrics: a Root Mean Square Error (RMSE) of 0.201, Mean Magnitude of Relative Error (MMRE) of 0.090, and Mean Absolute Percentage Error (MAPE) of 0.104.

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

Software Defect Prediction Using Machine Learning with Fractional Pufferfish Algorithm

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Gaurav Kishor Kanaujiya,
  • Prabhat Verma

摘要

Software defects are well known in software development and can create many issues for developers and users. Consequently, researchers have used distinctive methods to alleviate the implications of these flaws in source codes. Software defect prediction (SDP) provides observable results to development groups when contributing to industry decisions that predict defect code areas to support developers in identifying errors and arranging their testing activities. Thus, machine learning (ML)-based technique is designed for SDP. First, input software data are obtained and forwarded to the feature selection process. Feature selection is performed utilizing Information Gain (IG). The system forwards the selected features to the SDP process, where it applies the rider optimization algorithm-based neural network (RideNN), trained using the designed Fractional Pufferfish Optimization Algorithm (FPOA), to carry out SDP. The established FPOA was devised by incorporating Fractional Calculus (FC) and the Pufferfish Optimization Algorithm (POA). Finally, SDP analysis of SDP is performed with the learning data. Moreover, the FPOA-RideNN was evaluated using several metrics: a Root Mean Square Error (RMSE) of 0.201, Mean Magnitude of Relative Error (MMRE) of 0.090, and Mean Absolute Percentage Error (MAPE) of 0.104.