ML Algorithms: Geometric View

How it works geometrically:

Finds the best-fitting straight line through data points by minimizing the sum of squared distances (residuals) from points to the line. The line represents the relationship y = mx + b.

How it works geometrically:

Uses a sigmoid (S-shaped) curve to separate classes. The curve maps any input to a value between 0 and 1, representing probability. Decision boundary is typically at 0.5 probability.

How it works geometrically:

Creates rectangular decision regions by making binary splits parallel to feature axes. Each split divides the space into two regions, creating a tree-like hierarchy of decisions.

How it works geometrically:

Combines multiple decision trees, each creating different rectangular regions. The final decision is made by majority voting, creating smoother, more robust decision boundaries.

How it works geometrically:

Finds the optimal hyperplane that maximizes the margin between classes. Support vectors are the closest points to the decision boundary and define the separation.

How it works geometrically:

Uses Bayes' theorem with the naive assumption that features are independent given the class. Each class has its own (often Gaussian) distribution per feature; the overall likelihood is the product of 1D likelihoods. This leads to axis-aligned density contours and typically simple decision boundaries.