Datasets and Methods
The labeled and unlabeled dataset used to test the HEAL annotation server is available below.
The data consist of 9 features with numeric data type.
HEAL provides a reliable and costless alternative to existing Active learning that requires an human-in-the-loop to reinforce the accuracy of base model.