Incremental Decision Tree C++ Implementation

Question

Do anyone know any incremental implementation of decision tree classifier. Such that it could generate optimal decision tree classifier when you add new instance to training set with low computation and as quick as possible according existing decision tree classifier?

In other words I have an optimal decision tree classifier of set A, which named T_1, now I want to add instance X to set A and find optimal decision tree classifier tree T_2 by taking advantage of T_1 and X for set {A,X}.

adding instances will occurs several times. So it will valuable for me to find incremental method instead of building the tree every time.

Actually I am suspicious whether such implementation exists or not. It is appreciated if anybody could help me. I prefer code in c++ if there is any.

thanks

Answer 1

The wikipedia article links to two codes. IFI is not open source, but the source is avalable, VMFL is open source and in C (mostly)