Multiple Linear Least Squares Workbench

How to use Multiple Linear Least Squares Workbench

Predictor columns are centered by origin shifts and scaled to unit norm. Twice-reorthogonalized modified Gram-Schmidt supplies QR without forming normal equations. Back substitution finds slopes; centered residual and QR checks must pass before fitted rows or entered predictions are exported.

1. Enter predictor matrix json, response vector json, prediction matrix json (or []).
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

• 2–8 predictors, at least p+2 observations and at most 500 rows; up to 100 prediction rows. Numeric inputs have magnitude at most 1e9 and nonzero magnitude at least 1e-12. Constant or near-collinear columns with normalized QR diagonal at most 1e-8 are refused; coefficients above magnitude 1e15 are refused. Approximate arithmetic; the centered model is preferred when large origins cause an intercept to cancel. No p-values, confidence intervals, causal interpretation, regularization, missing data or extrapolation reliability claim.
• 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.