How namanhvu29/ss-ffgc-semi-supervised-fast-fuzzy-graph-cut works
SS-FFGC is a notebook-based semi-supervised clustering research project. Its documented workflow prepares datasets and labeled anchors, builds an RBF anchor graph, alternates cluster-center and soft-membership updates while fixing labeled rows, then compares clustering results and produces plots. The README describes these stages and outputs; notebook internals were not available to verify finer-grained calls.
Main entry points: SS_FFGC.ipynb Experiment Notebook · outputs Generated Results
Components 16
Data Preparation
- Experiment Notebook
SS_FFGC.ipynb - Benchmark Datasets
SS_FFGC.ipynb - Labeled Anchors
SS_FFGC.ipynb - COIL Features
SS_FFGC.ipynb
Clustering Algorithm
- RBF Anchor Graph
SS_FFGC.ipynb - Center Update
SS_FFGC.ipynb - Soft Membership Update
SS_FFGC.ipynb - Simplex Projection
SS_FFGC.ipynb - Hard Label Constraints
SS_FFGC.ipynb - Convergence Check
SS_FFGC.ipynb
Experiments
- Clustering Baselines
SS_FFGC.ipynb - Cluster Predictions
SS_FFGC.ipynb - ACC NMI ARI Timing
SS_FFGC.ipynb - Result Charts
SS_FFGC.ipynb - Generated Results
outputs
Other components
- Researcher
Connections 19
- Researcher → Experiment Notebook: runs experiments
- Experiment Notebook → Benchmark Datasets: prepares data
- Experiment Notebook → Labeled Anchors: selects labels
- Experiment Notebook → COIL Features: extracts features
- Labeled Anchors → RBF Anchor Graph: provides anchors
- Benchmark Datasets → RBF Anchor Graph: provides points
- COIL Features → RBF Anchor Graph: provides vectors
- RBF Anchor Graph → Center Update: provides graph
- Center Update → Soft Membership Update: updates centers
- Soft Membership Update → Simplex Projection: projects scores
- Simplex Projection → Hard Label Constraints: returns memberships
- Hard Label Constraints → Convergence Check: checks change
- Convergence Check → Center Update: repeats until converged
- Convergence Check → Cluster Predictions: yields memberships
- Experiment Notebook → Clustering Baselines: runs comparisons
- Cluster Predictions → ACC NMI ARI Timing: evaluates clusters
- Clustering Baselines → ACC NMI ARI Timing: compares results
- ACC NMI ARI Timing → Result Charts: summarizes results
- Result Charts → Generated Results: writes charts