Decision Tree Example: Reviewing the Tree

There are a number of options available to help tailor the building of a Decision Tree to your data.  These are explained in more detail in How do I set the Algorithm Options? and How do I set the Stopping Conditions?.

The Next Split panel allows you to investigate the decisions made at any particular split.  Each of the dimensions is examined and essentially a profile like report is generated showing the analysis and base counts within each value of the dimension (e.g. each income band).   See How do Decision Tree Statistics compare to profiles? for information on the differences between Profile and Decision Tree statistics.

The nodes produced at a split try to isolate customers in the analysis selection from those not in the analysis selection. 

The PWE algorithm below does this by separating into separate nodes, values which are above/below the Mean PWE for the variable.  This essentially isolates values for which the Analysis % is high, i.e. containing more Swedish holiday makers, from those where it is low.

The following example illustrates how the algorithm is applied:

The codes grouped together in branch 1 all have a PWE score greater than the Mean PWE (of -0.68).

  • The branch column is coloured red for branch 1, since the overall analysis % for this branch is higher than that in the root node

  • The individual categories are coloured according to their own analysis %.  As this decreases they are coloured less intense red, then grey and then light blue.

The codes grouped together in branch 2 all have a PWE score less than the Mean PWE (of -0.68).

  • The branch column is coloured blue for branch 2, since the overall analysis % for this branch is less than that in the root node.

  • The individual categories are coloured according to their own analysis %.  As this decreases they are coloured a more intense blue.

For more details on the Next Splits see the section Example using the Next Splits panel