I just happen to notice that there is a link between the SBT and the DFS(depth first search ) sequence. While training the model for the limited set of code say 10000 methods(due to computation limitation) the BLUE-4 Score looks for both SBT and DFS based model are very similar with the blue score difference less than 1. Is it due to my limited method number or is it due the similarity of the DFS and the SBT code.
I just happen to notice that there is a link between the SBT and the DFS(depth first search ) sequence. While training the model for the limited set of code say 10000 methods(due to computation limitation) the BLUE-4 Score looks for both SBT and DFS based model are very similar with the blue score difference less than 1. Is it due to my limited method number or is it due the similarity of the DFS and the SBT code.