<ul><li>Three factors - model architecture, dataset size, and class balance - influence test-time performance of machine-learning classifiers.</li><li>Evidence suggests dataset quality as an additional factor affecting classifier performance.</li><li>Study indicates quality is a dataset-intrinsic property, independent of model architecture, dataset size, and class balance.</li><li>Dataset quality is found to be an emergent property of the quality of datasets' constituent classes, providing a new target for optimization in machine-learning-based classification.</li></ul>

X-Factor: Quality Is a Dataset-Intrinsic Property

Discover more