Ganko

Ga
Ganko

Creator:Daniël Höhle
Function:Multi-dimensional decision tree algorithm
Input Type:xlsx, discrete
Input from Agent:Dizzy, as integer representation
Output Type:Set of decision rules, confusion matrix
Output to Agent:Advice, Ceres, Juno
Short Description:

Creates a multi-dimensional decision tree based on information gain (and their significance). It creates a number of root nodes and child nodes at once (the number is based on the maximum allowed and the minimum significance and information gain values). The branches of the tree are examined further on predictive value, where the best ones are used as a final prediction model - or as input to a following agent.





© 2015 Alan Turing Institute Almere