casics/spiral

Ronin splits

rsuhaibani opened this issue · 2 comments

rsuhaibani commented

The identifier [unloadAssemblies] has been divided as ['unload', 'Ass', 'embl', 'ies']

mhucka commented

Thanks for your report. Sometimes, some identifiers will inevitably be split suboptimally, and I guess this is an example of that. As mentioned in the README section on performance, the heuristic nature of Ronin makes it hard to be perfect for all cases. I think the only thing that could be done here is to retrain it on more/other examples or to come up with a different algorithm altogether.

  • 👍1
rsuhaibani commented

Clear! So, I can train it by myself on other examples and correct the results? I am working on analyzing more than 100K identifiers. Is it easy to do that?