Conceptual Question:
While attempting to match, I followed the phases: findTrainingData and label. However, when I proceeded to train, I encountered the following message:
24/12/06 12:16:31 WARN Client: Apologies for this message. Zingg has encountered an error. Unable to train as insufficient training data found. Training data has 14 matches and 1 non-match. Please run findTrainingData and label until you have sufficient labeled data to build the models.
Wouldn’t it be more efficient as a workflow to combine findTrainingData and label into a single integrated phase, iteratively identifying and labeling cases until the model gathers sufficient training data for downstream activities?
Streamlining the workflow by combining related phases (findTrainingData and label) into a more cohesive and iterative process. This would make it easier for users to collect sufficient training data without manually switching between phases, which can feel disjointed or repetitive.
