Models of Information and Sentiment Transmission (MIST)Transmission (MIST)

In counter-insurgency (COIN) and related operations (SSTR, HADR), it is critically important to understand the attitudes of different populations, but it remains difficult to get timely, detailed data for assessing mission effectiveness. Moreover, recent uprisings in the Arab world have highlighted the importance of early warning capabilities to detect areas of potential instability, quickly understanding emergent trends and how information spreads spatially and temporally. Paradoxically, the attitudes of populations globally are expressed online yet this readily available data has been vastly underutilized. To address these problems, Aptima proposes MIST (Models of Information and Sentiment Transmission). MIST leverages Aptima"s LaVA (Latent Variable Analysis) automated language processing technology to extract entities and high-level concepts from open source texts. In addition, our partners, SentiMetrix, Inc., will provide automated sentiment analysis on the same sources. These entities, sentiment, and concepts will be combined to derive"memes,"or units of cultural information about the attitudes and opinions present in the data. Finally, we will apply epidemiologically-inspired models of information diffusion to these memes to better understand their temporal and spatial change and spread. From these models we can derive both tactical and strategic metrics related to the changing attitudes of the population.