Open dpzhang opened 6 years ago
In paper Everyone’s an Influencer: Quantifying Influence on Twitter, it tries to compute the influence score for a given URL post. Hence, they need to track the origin of the URL diffusion. They use the time each URL posted. (see Section 4) To trace the origin of the manipulated emotions, I think what is harder is how to identify them. If the emotions transit gradually, we may need to set a threshold above which the manipulated emotion is discernible. My posted questions is also about the authenticity of the text being analyzed. 2018.1.8
Internet violence, negativity, manipulation of public opinions on the Internet have already been recognized as widespread epidemics in the age of social media. Kramer et al. used Facebook data and demonstrated how small changes in the visibility of positive and negative content within a user’s news feed would influence that user’s own posts. Thus, emotional contagion is likely to become a new form of systemic risks in the society, a byproduct of technological advancement and freedom of speech. In order to prevent intentional manipulation of this type of cascading effects for malicious purposes, I am wondering is it possible to identify manipulated emotions or attitudes? Is it possible to trace the origin of such emotions or attitude once a “spread” is detected? Is it possible to use computational tools to identify and to differentiate the authenticity of such emotions or attitudes by using corpuses of other text traces?