At the start of an active learning project, how can you estimate richness?

Estimating richness at the start of an active learning project involves assessing the potential value and diversity of the data being used. Using both methods mentioned allows for a more comprehensive and reliable estimation.

Running an early Project Validation can help identify key aspects of the project scope, data quality, and stakeholder requirements. This early assessment is crucial for determining whether the project's objectives align with the capabilities of the data set. It establishes a foundational understanding of what the project needs to achieve.

Coding a sample set of documents provides direct insight into the complexity and variability of the content. By reviewing a selection of documents, a project team can gauge how rich the data is in terms of relevance, context, and variability, which can affect the performance of the active learning model.

Combining both strategies enhances the estimation process. While Project Validation offers a strategic overview, coding documents delivers hands-on evaluation. By integrating both methodologies, a team can develop a well-rounded estimate of richness that informs how the project should proceed, such as whether additional data or specific resources are needed to achieve success. This dual approach maximizes the understanding of the project's context, leading to more informed decision-making as the project progresses.