A socially intelligent approach for enterprise information search and recommendation

Abstract

This paper focuses on the development of socially intelligent computing systems at the enterprise level. Specifically, in order to improve the information search and recommendation functionalities of social business software, we extend corporate knowledge structuring approaches, such as folksonomies and taxonomies, with the addition of statistical topic models. We use probabilistic models in order to uncover hidden topics in the corporate `knowledge base' and hence add an intelligent perspective in social collaboration. Probabilistic topic models are based upon the idea that documents are mixtures of topics, where a topic is defined as a probability distribution over words. We apply our approach in the Organik social business software platform and deploy it in five companies. Our results showed enhanced recommendations and improved search efficiency, while our approach effectively addresses problems in query expansion and recommends relevant resources and tags which in turn can leverage the creation and evolution of social knowledge structures like folksonomies.