Situation Tracking in Large Data Streams
Abstract
Work on this project resulted in a system which: (a) identifies in real-time a schema definition of a scenario from an incomplete sample of a situation(the cue) plus a corpus of real-time unstructured and sparse time series data (such as social media or log records); (b) uses this schema to identify additional implicitly relevant data records to provide much greater recall of the scenario data, and quantifies the presence of schema elements in each data record; (c) identifies multiple different scenarios emerging from the data in response to a query; (d) quantifies the strength and stability of the time series of data records which contribute to each unfolding scenario; (e) generates a user interface which enhances situation awareness and minimizes the user' need for accurate prior knowledge, by enhancing the user's search terms and collating results; and (f)employs sound data science to construct data stories from statistically valid information, so non-analysts can make good decisions.
Document Details
- Document Type
- Technical Report
- Publication Date
- Feb 01, 2015
- Accession Number
- ADA620924
Entities
People
- Andrew N. Smith
- Janet Wiles
Organizations
- University of Queensland