🧬 NEUROBLASTOMA SURVIVAL CLUSTERING ANALYSIS

This repository contains an unsupervised clustering analysis using Orange Data Mining to explore clinical profiles associated with survival in neuroblastoma patients.

🎯 Objective Identify patient subgroups based on biological and clinical features (e.g., MYCN status, risk level, tumor differentiation) and assess their correlation with survival outcomes.

📊 Dataset Derived from the study “Neuroblastomas in Eastern China” (PeerJ), the dataset includes 169 records with 12 features relevant to neuroblastoma prognosis.

⚙️ Methods Algorithms: Hierarchical Clustering, K-Means

Visualization: Box Plots, Bar Charts, Silhouette Score

Outcome variables (e.g., survival, follow-up time) were excluded from clustering and analyzed post hoc.

✅ Key Findings Clusters with high MYCN amplification, high-risk level, and low tumor differentiation were associated with lower survival rates. Longer follow-up times also aligned with better prognosis.