Synthetic Data for Design and Evaluation of Binary Classifiers in the Context of Bayesian Transfer Learning
This dataset is a collection of target and source synthetic datasets with different levels of relatedness for dimensions 2, 3, and 5. The classification complexity levels in the provided datasets are modeled through the Bayes error. For each pair of target and source datasets, the true model parameters are also provided. This dataset has been developed by: The Department of Electrical and Computer Engineering, Texas A&M University, College Station, TX 77840, USA & The Computational Science Initiative, Brookhaven National Laboratory, Upton, NY 11973, USA.