What is AdaBoost and why use decision stumps as weak learners?

AdaBoost builds a strong classifier from a weighted-majority vote of many weak learners. A decision stump — a single-feature threshold split — only needs to beat random guessing; AdaBoost's reweighting scheme forces each new stump to focus on the samples the ensemble currently gets wrong.

How is each round's best stump chosen?

The algorithm exhaustively scans candidate thresholds for every feature and both split polarities, computing the real weighted classification error under the current sample weights, and keeps the split with lowest weighted error.

How are sample and stump weights updated?

alpha = 0.5·ln((1-err)/err), scaled by the learning-rate control. Sample weights are multiplied by exp(-alpha·y·h(x)) and renormalised — the exact AdaBoost.M1 rule.

Home▸KI & Maschinelles Lernen▸Warenkorbabbruch-Vorhersage — Boosted Stumps live

🛒 Warenkorbabbruch-Vorhersage — Boosted Stumps live

Verfolgen Sie live, wie ein echtes AdaBoost-Ensemble aus Entscheidungsstümpfen anhand simulierter Shopping-Session-Merkmale trainiert wird, wobei jeder schwache Lerner die falsch klassifizierten Sitzungen der vorherigen Runde korrigiert.

KI & Maschinelles Lernen3DFortgeschritten60 FPS

About this simulation

This page runs a genuine AdaBoost (AdaBoost.M1) ensemble of decision-stump weak learners on a synthetic e-commerce dataset. Each simulated shopping session has five features — session duration, items in cart, cart value, page views and a returning-visitor flag — and a label generated from a real non-linear rule (a quick low-engagement bounce OR a high-value cart from a first-time visitor) plus roughly 8% label noise. Because no single axis-aligned stump can separate that rule, the ensemble's training error genuinely decreases only as more stumps combine.

Each boosting round performs an exhaustive weighted search over every feature, candidate threshold and split polarity to find the stump with lowest weighted classification error under the current per-sample weights, computes its vote weight (alpha) from that error rate, then exponentially upweights the samples it misclassified before renormalising all weights to sum to one — the exact textbook AdaBoost update.

Why do misclassified points visibly grow?

Each round's exponential weight update multiplies a misclassified sample's weight by e^(alpha), and the rendered point size is tied directly to that same weight — so a session the ensemble keeps getting wrong swells round after round until a future stump is finally forced to fix it.

Is the test accuracy real generalisation?

Yes — the held-out test set is generated from the same underlying rule but is never used to pick stumps or update weights, so its accuracy each round is a genuine measure of how well the growing ensemble generalises beyond the training sessions it was fit on.

⚙ Technische Details

Three.jsWebGLAIMachine Learning

3D · Three.js / WebGL renderer · 60 FPS target · runs fully client-side, no install