tkamal/ezehome-energy-intelligence

EZEHome Smart Home Energy Intelligence System

Trained models for the EZEHome Smart Housing Management platform.

Models

File	Algorithm	Test MAPE
xgboost.pkl	XGBoost (GridSearchCV)	2.32%
lstm.pkl	LSTM 64-unit (TF 2.17)	9.18%
kmeans.pkl	K-Means (k=7)	silhouette=0.49
iforest.pkl	Isolation Forest	contamination=0.01

Dataset

• 2-year synthetic Ezehome household data (Jan 2024 – Dec 2025)
• 17,520 hourly rows, 70/15/15 train/val/test split

Features

hour · day_of_week · month · is_weekend · lag_1h · lag_24h · lag_168h · rolling_mean_7d · rolling_mean_30d · delta_1h

Usage

from huggingface_hub import hf_hub_download
import joblib

model = joblib.load(hf_hub_download('tkamal/ezehome-energy-intelligence', 'xgboost.pkl'))