Agent-based simulation of twitter for building effective recommender system
Lubaid Ahmed, Abdolreza Abhari · Annual Simulation Symposium · 2014
This paper presents a framework to extract real-time information found in tweets using a multi-agent system for simulation of both a Twitter user and its attached agent. Agents communicate with each other and finally the framework is used to provide recommendation to its users. This process consists of the following tasks: extraction of distributed data using multi-agents from a social network website (i.e. Twitter), data cleansing, tweet analysis using Term Frequency/Inverse Document Frequency (TF/IDF), and providing recommendations. The aim of the proposed framework is to simulate Twitter users and analyze different information retrieval methods in a real-time distributed environment. Therefore, this simulator is capable of evaluating the performance metrics of the information retrieval methods together with scalability and distributed processing effectiveness. Simulation of Twitter users and reporting the scalability of information retrieval methods for processing the tweets are the new ideas presented in this paper.