---
license: mit
tags:
  - sklearn
  - solar-energy
  - time-series
  - regression
---

# Solar Power Forecast Model

RandomForestRegressor trained to predict plant-level DC power output
15 minutes ahead using weather sensor data and lag features.

**Dataset**: Kaggle Solar Power Generation Data (Plant 1, 34 days, 15-min intervals)
**Features**: irradiation, ambient temp, module temp, hour, day_of_year, month, lag_1, lag_4, rolling_mean_4
**R² (daytime)**: 0.9905
**R² (full dataset)**: 0.9323

## Usage

```python
import joblib
from huggingface_hub import hf_hub_download

path = hf_hub_download(repo_id="nakedved/genai-capstone", filename="solar_forecast_model.pkl")
model = joblib.load(path)
```