Machine Learning - Multi-response linear regression (Linear Decision trees)

Thomas Bayes

About

The multi-response linear regression method is a non-linear methods that build trees of linear models.

Representation

Model tree:

  • Each leaf has a linear regression model

How it will choose the node?

  • The Multi-response linear regression method will choose the class of an instance according to whichever of the three regression formulae produces the largest output. Ie the correlation coefficient is much smaller than for the others.

Non Linear Model Tree

Implementation

Weka

Tree > m5p





Discover More
Linear Vs True Regression Function
Machine Learning - Linear (Regression|Model)

Linear regression is a regression method (ie mathematical technique for predicting numeric outcome) based on the resolution of linear equation. This is a classical statistical method dating back more...



Share this page:
Follow us:
Task Runner