Explore core machine learning algorithms through interactive, client-side visualizations. Adjust hyperparameters, observe gradient trajectories step-by-step, and bridge the gap between mathematical rigor and visual intuition.
Predict continuous targets by fitting linear models, regularized loss functions, and polynomial curves.
Predict a numeric target by fitting the best straight line.
Fit a multidimensional hyperplane across multiple input features.
Shrink coefficients with an L2 penalty to guard against overfitting.
Analyze model complexity, underfitting, overfitting, and generalization error.
Separate data into discrete classes using probabilistic models, margins, partitions, and discriminant axes.
Predict a class probability from a linear decision boundary.
Find the boundary with the widest margin between classes.
Classify a point by a vote among its k closest labelled points.
Learn yes/no questions that partition data into pure regions.
Bayes’ rule with per-class bell curves and independent features.
Project data onto an axis maximizing between-class over within-class scatter.
Feed-forward multi-layer perceptrons learning non-linear representations via backpropagation.
Assess generalization error and avoid leakage through cross-validation and resampling.
Discover latent groupings, centroids, exemplars, and hierarchical structures in unlabeled data.
Split data into K groups around moving centroids.
Group data around actual member medoids (team captains) to resist extreme outliers.
Build clusters bottom-up or top-down with selectable linkage.
Grow clusters from dense regions and flag outliers as noise.
Compress high-dimensional spaces to extract principal components and visualize variance.