https://github.com/danieldacosta/energy-forecast
The repository consists of an energy forecasting model using XGboost. The dataset consists of hourly energy consumption rates in kWh for an industrial utility.
https://github.com/danieldacosta/energy-forecast
energy-forecasting-model forecasting xgboost
Last synced: about 1 year ago
JSON representation
The repository consists of an energy forecasting model using XGboost. The dataset consists of hourly energy consumption rates in kWh for an industrial utility.
- Host: GitHub
- URL: https://github.com/danieldacosta/energy-forecast
- Owner: DanielDaCosta
- Created: 2020-07-14T02:09:01.000Z (almost 6 years ago)
- Default Branch: master
- Last Pushed: 2023-05-01T13:43:55.000Z (about 3 years ago)
- Last Synced: 2025-04-06T03:41:15.733Z (about 1 year ago)
- Topics: energy-forecasting-model, forecasting, xgboost
- Language: Jupyter Notebook
- Homepage:
- Size: 1.32 MB
- Stars: 9
- Watchers: 0
- Forks: 2
- Open Issues: 1
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
# Energy Forecasting
This repository code came from the work of [Mario Dagrada](https://github.com/madagra/energy-ts-analysis), special thanks for the help.
The repository consists of an energy forecasting model using XGboost. The dataset consists of hourly energy consumption rates in kWh for an industrial utility over a period of around 7 months, from July 2019 to January 2020.
The final model has an *forecasting horizon* (The number of time periods to forecast into the future) of 48 time periods which corresponds to 2 days ahead forecasting.
# Details
Main techniques and terms used:
- Trend, Seasonality
- Stationary and Non-Stationary Time Series
- Augmented Dickey-Fuller Test
- Autocorrelation(ACF) and Partial Autocorrelation(PACF)
- Feature Engineering (lag features, standard time series features, endogenous features)
- XGBoost model with Bayesan hyperparameters optimization
The final model can be found inside ```model/```
# Usage
The Jupyter-Notebook, dataset and model were saved as Docker Image in GitHub Package.
You can install the package locally:
```
docker pull docker.pkg.github.com/danieldacosta/energy-forecast/energy-forecast_notebook:v1
```
You can also download it direct from the repository:
```
git clone https://github.com/DanielDaCosta/energy-forecast.git
```
# Acknowlegments & References
Special thanks to Mario Dagrada [Medium post](https://towardsdatascience.com/ml-time-series-forecasting-the-right-way-cbf3678845ff).
**References**:
- https://www.youtube.com/watch?v=Nm7m92sZZJA
- https://towardsdatascience.com/containerize-your-whole-data-science-environment-or-anything-you-want-with-docker-compose-e962b8ce8ce5