Label Error Detection

Label Error Detection

Datasaur's Label Error Detection feature, automating the identification of inaccuracies in datasets to enhance data integrity and model performance. Manual data review is time-consuming and prone to errors. With automated error detection, efficiency is boosted by pinpointing errors, focus is enhanced by applying specific error thresholds, and accuracy is improved with intelligent label suggestions.

Label Error Detection

To effectively use the Label Error Detection feature, make sure you have labeled your data and then you can start by enabling the extension from the Manage Extensions inside the project. Next, configure settings like the target column for input text and the target question for labels. Once configured, initiate the detection process and review the detected errors in the Label Errors section. Finally, adjust the error possibility threshold to refine your focus during the review process. See here for more detailed steps.