Are you a compatible user? Compatibility of a microblog user with a news article

Abstract

A novel concept of compatibility between a news article and an online microblog user is introduced, and a framework embodying the concept is proposed. The framework currently proposes to match two factors – user’s interest and user’s sentiment as reflected in the user’s microblog texts – to determine the compatibility. Using Twitter as an example, the framework is instantiated using the RAKE algorithm for topic keyword (for interest) matching and the VADER model-based sen-timent scoring algorithm for sentiment matching. Gold standard tests show that considering both interest match and sentiment match improves the accuracy of compatibility decision significantly and that filtering topic keywords based on co-occurrence semantics helps to disambiguate the user’s sentiment match, hence the compatibility decision.