Home▸AI & Machine Learning▸Data Preprocessing Hyperparameters

🧮 Data Preprocessing Hyperparameters

Watch a raw 3D point cloud of feature data reshape live as you switch between min-max, z-score and robust scaling, and dial in outliers and skew.

AI & Machine Learning3DAdvanced60 FPS

A raw 3D scatter of feature data reshapes live as you switch preprocessing methods, showing how min-max, z-score, and robust scaling each handle skew and outliers differently.

🔬 What It Demonstrates

Scaling method, outlier share, and feature skew are all preprocessing hyperparameters. The scene recomputes each feature's min/max, mean/std, and median/IQR live and repositions every point accordingly.

🎮 How to Use

Pick a scaling method, then push outlier share and skew up and down to see how each transform stretches, compresses, or stays stable. Resample for a fresh random dataset.

💡 Did You Know?

Robust scaling was designed specifically because min-max and z-score both anchor on statistics — the min/max or the mean/std — that a handful of extreme outliers can single-handedly distort.

⚙ Under the hood

Watch a raw 3D point cloud of feature data reshape live as you switch between min-max, z-score and robust scaling, and dial in outliers and skew.

машинне навчанняаналіз данихобробка сигналівалгоритмирозподіл данихінженеріяThree.js

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