Investigating Role of Deep Learning in Feature Selection for Fault Prediction Model in CBSE System

The focus of study is on defect prediction in software systems built from component parts. It has been suggested that the KC1 dataset could be useful in achieving this goal. Reusable software component assessment criteria are laid forth in this study. Research paper reviewed existing mechanisms that are applied of various dataset. In this study, the results for software defects have been presented when the datasets of CM1, KC1, KC2, KC3, MC2 and PC1 to PC3 have been taken into consideration and approaches such as J48, NB1, RF and CFS+NB have been utilized. A hybrid PSO-MVO optimization method might be used on a dataset of component selection factors. Provide intelligent and optimal solutions for component selection is the primary goal. A deep learning model is used to train on data that has been filtered after an optimal value has been detected. In order to assess the precision of the optimized component selection model, accuracy measures such as recall value, precision, and F1-score are taken into account. This kind of investigation is expected to have a substantial impact on the CBSE curriculum by delivering solutions that are both high-performing and accurate. Line count, cyclomatic complexity, design complexity, and estimated time, difficulty, intelligence, and effort all have their optimal values determined. The dataset might be filtered to develop an LSTM-based model to identify errors, taking into account the optimal value in future. Increased model dependability has resulted from the removal of non-optimized datasets and the selection of important features.