University of Tehran at RepLab 2014
Abolfazl AleAhmad, Payam Karisani, Maseud Rahgozar, Farhad Oroumchian · 2014
Abstract. In this paper, we present our approach to author ranking subtask; which is a part of author-profiling task in RepLab 2014. In this subtask, systems are expected to detect influential authors and opinion makers on Twitter web-site. The systems ’ output, for a given domain, must be a ranked list of authors according to their probability of being an influential author or opinion maker. Our system utilizes a Time-sensitive Voting algorithm, which is based on the hypothesis that influential authors tweet actively about topics of their interest. In this method, hot topics of each domain are extracted and a time-sensitive vot-ing algorithm ranks each authors on their respective topics.