vramados/house-price-model

House Price Prediction Model

PyTorch MLP trained on synthetic US housing data.

Model details

• —Architecture: Dense MLP (6 → 256 → 128 → 64 → 1)
• —Features: zipcode, housesqft, lotsqft, numbeds, numbaths, property_type
• —R2 Score: 0.9856
• —RMSE: ±$110,504

How to use

pip install torch scikit-learn python predict.py --zipcode 94102 --house_sqft 1500 ...

Training data

Synthetic US housing data — 1,000 rows, 10 cities, 50 zip codes.