Statistics Toolbox

Nonlinear Regression Models

Introduction

Not all relationships are well described by a linear regression model. The Statistics Toolbox provides two nonlinear regression techniques that may be preferable to linear regression in some cases.

• Nonlinear Least Squares fits a model that has a known parametric form but unknown parameter values.
• Regression and Classification Trees approximates a regression relationship using a decision tree. Such a tree seeks to partition the data set into regions, using values of the predictor variables, so that the response variables are roughly constant in each region.

Friedman's Test

Nonlinear Least Squares