SuraviAkhter/c-java-source-code

1---2annotations_creators:3  - expert-generated4language:5  - en6license: cc-by-4.07pretty_name: Cross-Project Defect Prediction (CPDP) Dataset — C & Java Projects8size_categories:9  - 1M<n<10M10source_datasets:11  - original12tags:13  - source-code14  - software-engineering15  - defect-prediction16  - transfer-learning17  - static-analysis18  - c-language19  - java20task_categories:21  - tabular-classification22task_ids:23  - multi-class-classification24---25 26 27 28# 🧩 Cross-Project Defect Prediction (CPDP) Dataset — C & Java Projects29 30This repository hosts a **custom dataset** for **Cross-Project Defect Prediction (CPDP)** research, curated from a diverse collection of real-world open-source projects written in **C (441 projects)** and **Java (98 projects)**.  31The dataset aims to support research on **software defect prediction, transfer learning**, and **imbalanced data handling** across heterogeneous programming environments.32 33---34 35## 📘 Overview36 37| Language | #Projects | Description |38|-----------|------------|-------------|39| **C** | 441 | Includes diverse open-source repositories from various domains |40| **Java** | 98 | Covers projects from academic domains collected from GitHub and other public repositories |41 42Each project folder typically includes:43- Source code files (`.c`, `.h`, `.java`)44- Bug/defect labels (where available)45- Metadata (e.g., LOC, complexity, commits)46- Preprocessed CSV feature files for ML models47 48---49 50## 🎯 Purpose51 52The dataset is designed for:53- **Cross-Project Defect Prediction (CPDP)**54- **Transfer Learning** and **Domain Adaptation** studies55- **Feature engineering** on static code metrics56- **Benchmarking** new software defect prediction models57 58---59 60## 🧠 Suggested Research Directions61 62- Comparison of **within-project vs cross-project** prediction accuracy  63- Study of **language heterogeneity** in CPDP (C ↔ Java transfer)  64- Use of **oversampling** or **class balancing** methods (e.g., SMOTE, OTOMO)  65- Integration with **Bayesian Networks**, **Deep Learning**, or **Tensor-based** models66