Explanations

These explanations dive into the what, why and theory behind each of DataEval’s features.

• Bayes Error Rate Estimation

• Cluster Assignment

• Detecting Bias in Datasets

• Cleaning Datasets

• Evaluating Metadata Parity

• Dataset Divergence

• Detecting Drift in Datasets

• Detecting Duplicates

• ImageStats and ChannelStats Classes

• Linter Class

• Model Training Techniques

• Detecting Out of Distribution Data

• Sufficiency of Datasets per Model

• Upperbound on Average Precision Estimation