A human is often only capable of labeling at 80% accuracy which leads to bad data. Theres a myriad of factors that can affect data quality and the only thing that's certain is that data quality can always be improved.

With this reality in mind we’ve created a feature which we call model corrects. The model looks at your data and then outputs what it has trouble understanding actually is a member of the class its said to be.

This allows you to clean up your data giving you better data quality for model training, decision making and customer focused implementations with your datasets.