ENSEMBLE-EMPOWERED TF-IDF: SCALABLE NLP FOR AUTOMATED ECLIPSE BUG REPORT CLASSIFICATION

Software maintenance in large open-source ecosystems depends heavily on fast and accurate bug triage. The Eclipse project alone receives tens of thousands of issue reports annually, covering a wide range of components and severity levels. Traditionally, human experts perform this triage manually, a process that is time-consuming, inconsistent, and increasingly unmanageable as the scale of the project grows. Previous attempts to automate this task using single machine learning models like Support Vector Machines (SVM) or Logistic Regression have achieved moderate accuracy (approximately 70–75%), but these approaches often rely on extensive feature engineering and struggle to generalize across evolving bug datasets. This research proposes a scalable, ensemble-based framework for automating Eclipse bug classification with high accuracy. The system processes raw bug descriptions through a comprehensive preprocessing pipeline—tokenization, stop word removal, and lemmatization—and converts the text into numerical representations using Term FrequencyInverse Document Frequency (TF-IDF). It then trains and evaluates five classifiers on a curated Eclipse–Mozilla dataset: SVM, Random Forest Classifier (RFC), Logistic Regression Classifier (LRC), Extra Trees Voting (EV) ensemble, and Extreme Gradient Boosting (XGBoost). A userfriendly GUI is integrated into the system, enabling non-experts to upload data, visualize preprocessing steps, and select models. With a 70/30 train-test split, the models yield the following performance: SVM achieves 74.23% accuracy, 83.03% precision, 73.94% recall, and 75.24% F₁score; RFC scores 83.51% accuracy, 87.68% precision, 83.40% recall, and 84.09% F₁; LRC records 70.10% accuracy, 75.06% precision, 70.19% recall, and 71.10% F₁; EV ensemble achieves 89.69% accuracy, 91.10% precision, 90.29% recall, and 90.21% F₁; while XGBoost outperforms all others with 92.27% accuracy, 92.91% precision, 92.65% recall, and 92.50% F₁-score. These results underscore the strength of ensemble methods, particularly XGBoost, in delivering reliable and scalable bug classification for large open-source projects.