In power system operations and electricity markets, missing data is a common problem in practice. This issue is especially significant when large-scale data-driven methods are used for point or probabilistic wind power forecasting. Data imputation methods, such as k-nearest neighbors and factor models, are crucial for filling in missing values before training forecasting models. These techniques ensure data completeness, which is essential for the accuracy of data-driven forecasting approaches.
-
Install Poetry
pip install poetry
-
Install Project Dependencies
Install the project dependencies, including MLflow, by running the following command in your project directory:
poetry install
-
Running the Experiments
To run the
experiment.pyscript within the Poetry environment, use the following command:poetry run python experiment.py
-
Viewing MLflow Tracking:
mlflow ui
After starting the MLflow UI, open your browser and go to http://127.0.0.1:5000 to view experiment results, including parameters, metrics, and artifacts.