Applied Mathematics & Information Sciences

Author Country (or Countries)



Business decision making has become complex due to increased online e-commerce activities (purchases). Though profitable, this has reduced personalized interactions with buyers, hence leaving the organizations in the leeward side of user’s feedbacks and requirements. Social networking sites provide a base platform for users to interact with their peers, also providing a platform for organizations to leverage the missed feedbacks and requirements. However, the huge amount of content (less relevant and more irrelevant) present in such domains makes appropriate data retrieval a complex task. This paper presents an architecture that can be used to effectively retrieve information from heterogeneous data sources based upon product based queries presented by user. Major root causes related to the users query are identified from the retrieved information. The root causes are segregated in terms of their polarity, hence providing results of higher significance. The identified root causes can be used for effective decision making. Efficiency of the retrieval levels and sentiment prediction levels are experimentally evaluated and were found to be effective in terms of scalability, retrieval levels and accuracy levels

Digital Object Identifier (DOI)