What is the maximum number of documents recommended to be included in a saved search for Active Learning?

The recommended maximum number of documents to include in a saved search for Active Learning is set at 9 million. This limit is based on the balance required to ensure efficient processing, training, and retrieval of relevant documents while maintaining a high level of effectiveness in the machine learning algorithms.

While larger datasets can provide more information for the model to learn from, excessively high numbers beyond this threshold can lead to diminishing returns, increased computational load, and longer processing times. Therefore, setting a cap at 9 million allows for a significant yet manageable dataset that can enhance the accurateness of predictive coding without overwhelming the system's capabilities.

Additionally, adhering to this maximum promotes a focus on quality over quantity, which is crucial in machine learning environments where the accuracy and relevance of the data directly affect the outcomes of the Active Learning process.