Fill gaps from assistant conversations: scalers, SES alpha, fitted_

[mintlify]

Summary

Fill three recurring documentation gaps surfaced by the assistant conversation signal so readers can answer them without help.

Changes

• neuralforecast/docs/capabilities/time_series_scaling.html.mdx: distinguish robust (median + MAD) from robust-iqr (median + IQR), fix the option list typo (minmax1 instead of duplicated minmax), and link to LocalRobustScaler for the implementation.
• statsforecast/docs/models/simpleexponentialsmoothing.html.mdx: note that the fixed alpha is applied uniformly to every series and point to SimpleExponentialSmoothingOptimized for per-series smoothing.
• statsforecast/docs/models/simpleexponentialoptimized.html.mdx: state that alpha is optimized independently per unique_id and there is no way to bound the search range.
• statsforecast/src/core/core.html.mdx: document the sf.fitted_ array shape (n_series, n_models), the ordering of each axis, and show how to iterate over series and access residuals via .model_.

[mintlify]

Preview deployment for your docs. Learn more about Mintlify Previews.

Project	Status	Preview	Updated (UTC)
nixtlaverse_nixtla_io	🟢 Ready	View Preview	Aug 24, 2026, 4:15 PM

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}