KBSI is researching and designing technology that allows users to rapidly build text analytics-based systems that support the research and analysis of large bodies of disparate, unstructured text sources.
The Text Analytics Situational Awareness Toolkit (T-SAT) will apply an advanced object model that allows users to combine multiple text analytics algorithms to create complex research and analysis scenarios. The technology will provide a user configurable GUI in which users can rapidly construct research and analysis systems that connect data sources to analytics-based algorithms. T-SAT’s text analytic-based algorithms can be applied to a wide range of source types–unstructured text from an online social network, text from an online chat room, or digitized textual reports taken from human intelligence sources (HUMINT)–eliminating the need for tailoring algorithms according to each target source.
In addition, the T-SAT object model will provide creative flexibility in how systems are constructed. Users can construct a T-SAT system from scratch or by dragging and dropping components from the T-SAT library into the management GUI and then connecting and arranging them as needed. Analysts will be able to connect data sources with analytics-based algorithms using models capable of recognizing certain signature profiles: anti-West sentiments, indicators of deteriorating social, political, or economic conditions, etc.
This initial source list can be filtered in T-SAT and connected to additional text analytics-based algorithms capable of extracting social network models and activity profiles. The T-SAT system can be configured to monitor these data sources over time, updating social networks by tracking their dynamics and generating alerts of any threatening activity signatures.
When a user is satisfied with a constructed model, he can compile the model into a computational footprint that can then deployed and executed. T-SAT will support text analytics-based system deployment to small portable devices and large-scale, multi-processor platforms such as Hadoop and other cloud-based architectures. Users can save their models, adding them to the library of T-SAT models. Archived models can be used in building new models or can be combined with existing models.
The T-SAT system will support a fluid and creative environment for data and intelligence analysts in which the burden of data source, text analytics-based algorithms, and visualization details are removed. This gives analysts more time and energy to expend on data exploratory efforts rather than on implementation headaches. These innovative situation awareness capabilities will improve analysts’ contextual understanding of data sets and significantly reduce all important ‘data-to-decision’ times.