Abstract
In recent years, Online Social Networks (OSNs) have attracted a significant increased number of users. New methods for extracting data are required to deal with the real time changes of a huge amount of personal information in OSNs. In the past, we implemented a parser as centralized system to retrieve information from OSN profiles source web pages. One of the drawbacks was that the parser had to be updated to reflect the changes in the profiles' structure. In this paper, we extend our previous work that proposed Online Social Network Retrieval System (OSNRS) to decentralize the retrieving information process from OSN. The novelty of OSNRS, which is based on a Multi Agent System (MAS), is its ability to monitor profiles continuously. The new addition involves replacing the parser with the Application Programming Interface (API) tool to enable OSNRS to be integrated with services that are supported by OSN providers in the absence of the profiles source web page. Also, new algorithms alongside case studies are presented to improve OSNRS. The experimental work shows that using API and MAS simplifies and speeds up tracking the history of OSN profiles. Moreover, combining them with text mining helps us further to understand the dynamic behaviour of OSNs users.