Hi Chintan, 

Do you mean you are trying to improve Decision Tree itself through modifying its core concept or you want to improve the result of Decision Tree when applying it to classify a particular dataset??

Please make your question clear then we can help you.

Kind regards

Ahmed

Hi,

You can use random forest instead decision tree.

If the dataset is really "massive", then you can use online decision trees or online random forests.

See for instance VFDT:

http://homes.cs.washington.edu/~pedrod/papers/kdd00.pdf