A step-by-step walkthrough of how a decision tree picks where to split a node. With five observations and four candidate thresholds, the artifact traces the RSS computation for each split and lets the reader see why greedy search gains in clarity what it loses in global optimality. Interactive animation and live computation, bilingual.
Interactive artifact
Related artifacts
The Tradeoff: Bias and Variance
An interactive scrollytelling artifact explaining the bias-variance tradeoff through the U-shaped test-error curve from ISLR Chapter 2. Six visual steps: from raw data to cross-validation, rendered in a sumi-e ink-on-washi aesthetic. Bilingual: Spanish and English.
GLM for Actuarial Models
From theory to production: how insurers actually estimate claim frequency, severity, and policy lapse. This note walks from exponential families and link functions through validation with ROC curves, Gini coefficients, and relativity tables. Three full case studies: auto, life, and workers' comp. Personal actuarial synthesis note.