Classic machine learning algorithms and methods, from regression and decision trees to clustering, ensembles and feature engineering.
56 terms
SMOTE is a machine learning resampling method that creates synthetic minority-class samples by interpolating between nearby real examples.
Spectral clustering groups data by building a similarity graph and splitting it with the eigenvectors of the graph Laplacian, finding non-convex clusters.
Stacking in machine learning combines several base models by training a meta-learner on their out-of-fold predictions to improve accuracy.
A support vector machine (SVM) is a supervised machine learning model that separates classes with the widest possible margin between them.
t-SNE is a nonlinear machine learning technique that maps high-dimensional data to two or three dimensions while preserving local neighborhoods.
UMAP is a nonlinear dimensionality reduction method in machine learning that maps high-dimensional data to a few dimensions while keeping local structure.