How hzavadil98/uit_fys-8603 works
This is a research codebase combining breast-cancer image classification experiments, synthetic image data, and multiview statistical analysis. The intended ML workflow is to prepare views, train a model, and produce predictions and evaluation metrics; the tree identifies training entrypoints, but their wiring was not sampled, so those links are omitted. The statistical packages provide separate PCA, MCCA, and AJIVE analyses. There is no README, and unsampled modules are represented only where their paths and roles are clear from the tree.
Main entry points: train_synt.py Synthetic training · Synthetic_2v.py Two-view image dataset · train_elements.py Element training · ajive_fun.py AJIVE analysis
Components 24
Cancer ML
- Two-view training
train_2v1b.py - Four-view training
train_4v2b.py - Feature training
train_4vf.py - Breast cancer data
src/Dataloaders/Breast_Cancer.py - Patient cancer data
src/Dataloaders/Patient_Cancer.py - View cancer data
src/Dataloaders/View_Cancer.py - Two-view models
src/Models/Two_view_grey_models.py - Separate-view models
src/Models/Separate_view_models.py - Multiview cancer models
src/Models/Multiview_cancer_models.py - MIRAI pretrained model
src/Models/mirai_pretrained.py
Synthetic Data
- Synthetic training
train_synt.py - Two-view image dataset
src/Dataloaders/Synthetic_2v.py - Element datasets
elements/elements/classes.py - Element training
train_elements.py - Synthetic data generation
src/Dataloaders/Synthetic_2v.py
Multiview Analysis
- AJIVE analysis
python_packages/mvdr/ajive/ajive_fun.py - MCCA analysis
python_packages/mvdr/mcca/mcca.py - JIVE analysis
python_packages/jive/AJIVE.py - DIVAS bridge
src/py_divas.py
Statistical Tools
- PCA
python_packages/ya_pca/PCA.py - Rank selection
python_packages/ya_pca/rank_selection/rank_selection.py - Subspace comparisons
python_packages/mvdr/principal_angles.py
Other components
- Researcher
- R DIVAS package
Connections 16
- Researcher → Two-view training: runs
- Researcher → Four-view training: runs
- Researcher → Feature training: runs
- Researcher → Synthetic training: runs
- Researcher → Element training: runs
- Researcher → AJIVE analysis: analyzes views
- Researcher → MCCA analysis: analyzes views
- Researcher → JIVE analysis: analyzes views
- Researcher → PCA: fits components
- Researcher → DIVAS bridge: runs analysis
- DIVAS bridge → R DIVAS package: calls
- Synthetic training → Two-view image dataset: uses dataset
- Two-view image dataset → Synthetic data generation: generates images
- Element training → Element datasets: uses datasets
- PCA → Rank selection: selects rank
- AJIVE analysis → Subspace comparisons: compares subspaces