Using LT to introduce errors

I would like create a dataset with some grammatical errors, similar to those detectable by LT.

Is it possible to use LT to perform this operation, e.g., by reverse applying rules on clean english sentences.

Thank you.

Yes, you will need to create XML pattern rules that introduce errors.

Then you can run, only with these error-introducing rules enabled, and with this option:

-a, --apply automatically apply suggestions if available, printing result to STDOUT
      NOTE: only use with very robust rules, as this will otherwise introduce new errors

Hi @jaumeortola
Thank you for your reply.
I don’t have XML patterns. Is it possible to create such rules automatically from existing correction rules, e.g. by “inverting them” ?