OPTIMIZING SOFTWARE DEFECT PREDICTION USING A HYBRID NEURAL NETWORK WITH WEIGHTED MODIFIEDCUCKOO SEARCH

Flaws significantly impact the quality, dependability, and maintenance of software.Discovering flaws before they appear in production environments is possible using data mining and classification algorithms.However, precise prediction necessitates managing intricate datasets.HNN with WMCSO, a soft computing model, is proposed in this study to identify software defects.Before extracting the noteworthy new features from the dataset, the proposed model uses the MFCM to perform the clustering procedure.The HNN is utilized for both software fault classification and prediction, and its weights are adjusted using the WMCSO model.Prediction rate and execution time analyses are used to rate the HNN-WMCSO methodology.In terms of an effective prediction rate, the experimental inspection indicates that the suggested model works better than the current approach.