Discover Stressor: An R package created to help researchers determine the accuracy of predictive models quickly. The package allows R users to leverage the power of Python's PyCaret library without learning Python itself, making it easier to compare and refine machine learning models. With synthetic data set generation functions and accuracy comparison metrics, this package is a valuable tool for machine learning experts and beginners alike.
This thesis investigates Machine Learning models in asymptotic regimes. Researchers have found contradictory results regarding traditional statistical wisdom. It uses LASSO & Random Features regression to figure out why. Gaussian comparison theorems expose these models’ properties, including overparameterization & lower order statistical moments to determine universality. This research promises to help solve Machine Learning mysteries.