Gender Discriminatory Language Identification with an Hybrid Algorithm based on Syntactic Rules and Machine Learning

In the last years, gender discrimination in textual documents has emerged as an open problem and is undergoing analysis. The difficulty of identifying sentences in which this discrimination is present is linked to the context used and the formalisms adopted. This work describes an exploratory activity linked to the context of regulations and official documents of Italian public administrations. A hybrid algorithm based on syntactic rules and machine learning is therefore proposed, capable of identifying a specific subset of possible gender discrimination.