This repository contains the code and data related to the paper on Categorization of Comparative Sentences at the ACL 2019 workshop on Argument Mining. This paper presents a set of machine learning classifiers which perform categorization of sentences into three classes: BETTER, WORSE, or NONE wehere each sentence is expected to contain mentions of two objects under a comparison. For instance, consider the following sentence:
Python is better than Ruby for scientific programming. In this sentence, Python is compared to Ruby with respect to the aspect "scientific programming". Our classifier is expected to categorize it with the BETTER label as the first mentioned object is better than the second mentioned objext. Based on such comparisons one can compare objects in general.
This repository contains also the result of crowdsourcing annotation campaign to create a training dataset, and the result of serveral classification experiments. In the experiments we rely on some libraries: