CAM checks wether a sentence like "Python is better than Ruby" contains a comparison and, if it does, which object is the prefered one.
This repository contains the result of crowdsourcing annotation campaign to create a CAM data set, and the result of serveral classification experiments.
A paper based on the thesis is available at ArXiv (eprint arXiv:1809.06152)