The Caltech Data Engine is a computer program that contains several predefined data models, such as neural networks, support vector machines (SVM), and radial basis functions (RBF). When requested for data, it randomly picks a model, generates (also randomly) parameters for that model, and produces random examples according to the generated model. A complexity factor can be specified which controls the complexity of the generated model. The engine can be prompted repeatedly to generate independent data sets from the same model to achieve small error bars in testing and comparing learning algorithms.
The DEngin was developed by Amrit Pratap.
The Caltech Data Engine has been used in several courses and papers.