AutoML depends on the Python packages imbalanced-learn and scikit-learn. Use the following command to install these packages before running AutoML.
pip install teradataml[automl]
AutoCluster leverages open-source machine learning functions for model training, which requires the appropriate cluster configuration to ensure full support of these capabilities. For detailed guidance, see teradataml OpenSourceML Setup and Requirements.
Pandas version support
AutoML is currently supported with pandas 2.x only.
pandas 3.0 and above are not supported at this time.
To ensure successful execution of AutoML workflows, use:
pandas >= 2.0 and < 3.0
scikit-learn version support
Feature selection using Lasso is currently supported up to scikit-learn 1.7.2.
For scikit-learn 1.8 and above, you must either disable Lasso-based feature selection or downgrade scikit-learn to version 1.7.2 to run AutoML successfully.