VFDT reduces training time for large incremental data sets for the sub sampling incoming data stream. CVFDT can adapt to concept drift, by using a sliding window on incoming data. Old data outside the window is forgotten. VFDTC extends VFDT for continuous data, concept drift, and application of Naive Bayes classifiers in the leaves. I've collected this data from wiki. You can find some useful information form the attached paper. 

Hi,

I would like to add, that there are some "new" and interesting problems, where almost the same algorithmic approach than VFDT can be applied (building online decision trees).

In particular the contextual bandit problem, where a player interacts with the environment by playing actions knowing contextual information, and updates its policy with the feedback provided by the environment.

This problem is used for instance to choose the ads on websites: each potential ad is an action, the context is the profile of the web page, of the user, the feedback is the click, and the policy is a model such as a decision tree.

For use cases see:

https://www.researchgate.net/publication/290104476_Bandit_Forest

For the algorithm and its analysis see:

https://www.researchgate.net/publication/289952125_Random_Forest_for_the_Contextual_Bandit_Problem_-_extended_version

Presentation Bandit Forest