Treballs Finals de Grau d'Enginyeria Informàtica, Facultat de Matemàtiques, Universitat de Barcelona, Any: 2016, Director: Maite López Sánchez i Juan Antonio Rodríguez-AguilarDiscussion has changed enormously over the last decades. Thanks to the internet and advancements on the field, we are able to have larger communities in a discussion and get higher quality results from it than ever before. Having a large number of individuals involved has lots of benefits, but it also carries some challenges. As a discussion grows larger, people have a tendency to start repeating arguments, this has been traditionally handled by a team of moderators. In this work we analyse the properties of these arguments, and we take advantage of those properties to make a specialized method to automatically detect syntactically different but semantically equivalent arguments. Thus reducing the amount of work carried out by moderators. We do so with the help of natural language processing and machine learning techniques.