The Taboo algorithm has the particularity of being able to start from an existing solution, and tries to improve it by allowing the deterioration of the score to get out of local optima.

New Feature: Post-Processing with Taboo

As of version 8.0, you can now choose to post-process with Taboo the network you have learned with:

  • Maximum Spanning Tree,
  • EQ,
  • SopLEQ, and
  • Taboo Order 

This automated workflow is available in Data Perturbation, Structural Coefficient Analysis, and can also be evaluated with Cross-Validation.