You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Comprehensive project workspace for software defect prediction experiments with explainable AI (XAI), cross-project transfer analysis, statistical testing, and academic reporting artifacts.
Master repository documentation and navigation guide.
XAI.docx
193.67 KB
2026-04-30 01:42
Project write-up / manuscript artifact in Word format.
.vscode/settings.json
36 B
2025-10-02 15:19
Editor workspace settings for VS Code.
Top-Level Directories
Directory
Files
Subdirectories
Directory README
dataset/
10
0
dataset/README.md
documentation/
9
0
documentation/README.md
monthly_reports/
8
0
monthly_reports/README.md
research_papers/
8
0
research_papers/README.md
results/
14
1
results/README.md
src/
1
6
src/README.md
src/ Notebook Breakdown
Directory
Notebook Count
Role
src/approaches/
4
Jupyter notebooks
src/within_current/
8
Jupyter notebooks
src/within_proposed/
8
Jupyter notebooks
src/cross_current/
13
Jupyter notebooks
src/cross_proposed/
8
Jupyter notebooks
src/testing/
2
Jupyter notebooks
Detailed Directory READMEs
dataset/README.md
documentation/README.md
monthly_reports/README.md
research_papers/README.md
results/README.md
results/zcross_plots_mean/README.md
src/README.md
src/approaches/README.md
src/cross_current/README.md
src/cross_proposed/README.md
src/testing/README.md
src/within_current/README.md
src/within_proposed/README.md
Usage Notes
Keep datasets under dataset/.
Keep generated numeric outputs and figures under results/.
Keep semester and defense material under documentation/ and monthly_reports/.
Keep all experiment execution notebooks under src/.
Maintenance Rule
Whenever files are added, renamed, or removed in any documented folder, update that folder's README in the same commit to keep this repository self-describing.
About
TrustX-Defect is an explainable AI pipeline that predicts software defects from tabular datasets and pairs strong model performance with transparent SHAP-based insights. It automates preprocessing, cross-validated training, and reliability analysis so teams can trust and interpret their defect forecasts.