Weighted Isotonic Regression Workbench

How to use Weighted Isotonic Regression Workbench

Pool-adjacent-violators merges neighboring weighted-mean blocks whenever their order violates your selected direction. Exact rational arithmetic preserves ties and weighted means. The result verifies monotonicity through the block rule and cumulative weighted residual conditions.

1. Enter observation json (value, weight).
2. Choose the stated convention and check the units and supported model.
3. Calculate, inspect the result, then copy or download it.

Good to know

• Exact-grid inputs are plain decimals with at most six places and magnitude≤1e9. Raw JSON number spelling is checked before conversion; decimal strings can preserve exact grid values. Exact rational intermediates are bounded to 8192 bits. Up to 500 ordered observations with positive weights. Ordering is your row order, not a sorted x column. This is weighted squared-error projection onto a monotone sequence, not linear regression or a probability calibration guarantee. No extrapolation, confidence bands or causal interpretation.
• Local calculation only; no input persistence or external lookup. Text/JSON is limited to 30000 characters and output to 1000000 characters. Numeric model inputs and precision/work limits are stated on this page. Approximate JSON numbers retain full stored values; exact arithmetic outputs retain decimal or fraction strings. Unsupported states fail explicitly.