How to classify relationships (NLP)?

Question

data mining

Hello!
There is some unmarked corpus, triples of type (OBJECT, RELATIONSHIP, SUBJECT) are extracted from it. Relationships were extracted using Open Information Extraction methods. Of these triples, you need to choose only those that relate to financial events.
Only relationship clustering comes to mind, but the number of words in the relationship is too small. Has anyone encountered a similar problem

Answer 1

It may be worth marking (training) something to recognize "financial events".
For example Bayesian classifier: https://github.com/loguntsov/bayes