System and method for evaluating the credibility of news emerging in social networks for information and news reporting purposes

摘要

A system for the evaluation of the use of social networks for informational purposes. The system consists of a Natural Language Processing (NLP) engine, a network analytics engine, a storage layer capable of handling big-data and an integration sub-system which interconnects it with external social networks. The method includes the valuation of the social networks users and the validation of the content which is shared by them. A wide array of metrics is introduced to quantify the measured characteristic properties that are identified. The metrics are analyzed according to the measured pillar (contributor, content, context) and evaluated with the use of data structures internal to the system, as well as known NLP techniques (for classification, clustering, topics modeling, events identification, sentiment analysis etc.) and networked service requests (API calls). The calculated values for each metric are reduced to an index by fitting the data to the appropriate distribution (as calculated by the networks analytics engine), eventually resulting in an index value for each of the system metrics. The solution employs the most complete array of metrics compared to the state of the art and the processing it performs makes it particularly practical.