MoRS at SemEval-2017 Task 3: Easy to use SVM in Ranking Tasks

Proceedings of the 11th International Workshop on Semantic Evaluations (SemEval-2017), pages 287–291, Vancouver, Canada, August 3 - 4, 2017. c⃝2017 Association for Computational Linguistics MoRS At SemEval-2017 Task 3: Easy To Use SVM In Ranking Tasks Miguel J. Rodrigues∗1 and Francisco Couto1 1LaSIGE, Faculdade de Ciˆencias, Universidade de Lisboa, Portugal mrodrigues@lasige.di.fc.ul.pt, fcouto@di.fc.ul.pt Abstract This paper describes our system, dubbed MoRS (Modular Ranking System), pro- nounced ’Morse’, which participated in Task 3 of SemEval-2017. We used MoRS to perform the Community Question An- swering Task 3, which consisted on re- ordering a set of comments according to their usefulness in answering the ques- tion in the thread. This was made for a large collection of questions created by a user community. As for this challenge we wanted to go back to simple, easy-to-use, and somewhat forgotten technologies that we think, in the hands of non-expert peo- ple, could be reused in their own data sets. Some of our techniques included the anno- tation of text, the retrieval of meta-data for each comment, POS tagging and Named Entity Recognition, among others. These gave place to syntactical analysis and se- mantic measurements. Finally we show and discuss our results and the context of our approach, which is part of a more com- prehensive system in development, named MoQA. 1 Credits This work was supported by FCT through funding of LaSIGE Research Unit, ref. UID/CEC/00408/2013. 2 Introduction The main difference between Question Answering (QA) and Community Question Answering (cQA) is, while QA systems rely on a user query in or- der to search and prepare an answer based on the ∗Corresponding author: mrodrigues@lasige.di.fc.ul.pt searching capabilities it already has and its docu- ments, in cQA the query and respective related an- swers are already provided, being only necessary a reordering by relevance of such answers, or per- haps even a rephrasing of such an answer in order to suit better the query. In recent times, QA systems have attracted great interest in the information retrieval community, and also in the cQA (H¨offner et al., 2016). A characteristic of cQA, is that a user resorts to the web for answers without a given structured knowledge base. The arbitrariness of cQA forums, the dependence and waiting time on their results, may slow the gathering of answers in real time. Also, public forums are dependent of the users in- put (i.e. answers), which might be rather unstruc- tured, not straight to the point, not related to the question at hand, not well written (i.e. grammati- cally), lengthy or even incorrect. Our team shares this exact same interest and continuous develop- ment in this area. Our participation focused on the Community Question Answering (CQA) Task 3 SubTask A of 2017 SemEval edition 1 (Nakov et al., 2017), which consisted on reordering a set of comments according to their usefulness in answering the question in the thread. This was made for a large collection of questions created by a user com- munity, provided by the task organizers. Suc- cinctly, Subtask …

Content truncated to 500 words