User Guide

Comprehensive documentation for all LOESS parameters and features.

• Parameters

  ---

  All configuration options explained

  Parameters

• Execution Modes

  ---

  Batch, Streaming, and Online adapters

  Adapters

• Weight Functions

  ---

  Available kernel functions

  Kernels

• Robustness

  ---

  Outlier handling methods

  Robustness

• Intervals

  ---

  Confidence and prediction intervals

  Intervals

• Cross-Validation

  ---

  Automated parameter selection

  Cross-Validation

• Polynomial Degree

  ---

  Local constant, linear, and quadratic fits

  Degree

• Multivariate LOESS

  ---

  Smoothing over 2-D and n-D predictor spaces

  Dimensions

• Boundary Handling

  ---

  Edge strategies to reduce boundary bias

  Boundary

• Scaling Methods

  ---

  Residual scale estimation for robustness iterations

  Scaling

• Merge Strategies

  ---

  Reconciling overlapping chunks in Streaming mode

  Merge