linear-regression-tetuan-power

Model Name

linear-regression-tetuan-power — Linear Regression (Phase 1)

Algorithm

sklearn.linear_model.LinearRegression (ordinary least squares), trained with scikit-learn 1.6.1. No other regression algorithm is used in this model.

Training Information

• Dataset: Tetuan City power consumption dataset (development data only; the deployed model is not tied to this dataset's specific values, only its feature schema).
• Rows used for development: 44553 (of 52416 total, after validation/cleaning); 7863 rows were held out and never seen during training.
• Train/test split: 80% / 20%, random_state=42.
• Train samples: 35642
• Test samples: 8911

Input Format

A pandas DataFrame (or 2D array) with these numeric columns, in this order:

• Temperature
• Humidity
• Wind Speed
• general diffuse flows
• diffuse flows

Target Format

A single numeric column: Zone 1 Power Consumption (float, same units as the training target).

Evaluation Metrics

Computed on a held-out 20% test split from the development pool (never used in training):

Metric	Value
MAE	5119.2703
MSE	39281693.2639
RMSE	6267.5109
MAPE	0.1658
R²	0.207988
Adjusted R²	0.207543

How to Download the Model

from huggingface_hub import hf_hub_download

model_path = hf_hub_download(
    repo_id="YOUR_USERNAME/linear-regression-tetuan-power",
    filename="linear_regression.joblib",
)

How to Load the Model

import joblib

model = joblib.load(model_path)

Example Prediction

import pandas as pd

sample = pd.DataFrame([{
    'Temperature': 0.0,
    'Humidity': 0.0,
    'Wind Speed': 0.0,
    'general diffuse flows': 0.0,
    'diffuse flows': 0.0,
}])

prediction = model.predict(sample)
print(prediction)

Do not retrain this downloaded model — it is provided purely for inference. See test_downloaded_model.ipynb in this repository for a full download -> load -> predict -> evaluate example on a different CSV dataset.