Child pages
  • Multi-Quadrant (5.4)

Versions Compared

Key

  • This line was added.
  • This line was removed.
  • Formatting was changed.

...

New Feature: Structural Learning Algorithms

Prior to this release , it was only possible to estimate the parameters (5.4, Multi-Quadrant Analysis was limited to estimating the parameters of the current network, i.e. estimating marginal and conditional probability tables) of the current network on the different datasets , for the subsets of data defined by the states of the Breakout Variable. The structure was then identical for all the resulting networks.As a result, the structure of all networks would be the same. 

Now, you can choose to run any Unsupervised Structural learning algorithms, or any Supervised learning algorithms on , for each subset of your data.

...