Multi-lingual Social-media Crowd Manipulation Detector (MSCMD)

In this SBIR, BCL proposes developing a Multi-lingual Social-media Crowd Manipulation Detector (MSCMD). The MSCMD will use natural language processing techniques to detect terms that arouse emotion using information out of context to trigger reaction from the audience and move them to act.The MSCMD will operate in Asian languages using a Natural Language Processor for each language. The MSCMD will then parse the text and transforms them into a parse tree with parts of speech tagging. The parse tree will then be tagged for features that have high emotional content, and an emotional feature model will be applied to determine the type of action the text is attempting to raise.A machine learning model will then be applied to the collected data to develop adaptive models that can detect Crowd Manipulation.In Phase I, BCL will research the feasibility of using NLP techniques and psychological markers to detect Social Media Manipulation in the Asia-Pacific region.In Phase II, BCL will develop a machine learning engine that automatically learns what emotions are being used to Manipulate Social Media.