SeismoSafe Nepal – Earthquake Damage Classifier
LightGBM classifier that predicts building damage levels from the 2015 Gorkha earthquake in Nepal using structural, geographical, and ownership data.
Labels: 1 = low damage, 2 = medium damage, 3 = almost complete destruction
Metric: Micro-averaged F1 | Score: 0.7433
Usage
from huggingface_hub import hf_hub_download
import joblib
model = joblib.load(
hf_hub_download("un1u3/seismosafe-nepal", "sklearn_model.joblib")
)
preds = model.predict(X)
Model
- Algorithm: LightGBM
- Validation: 5-fold stratified CV
- Key features: Geo target encoding, material risk scores, structural interaction features
Author
Unique Shrestha — GitHub | uniquestha422@gmail.com
- Downloads last month
- -