A Framework for Analyst Focus from Computed Significance

Abstract

Attention is the critical resource for intelligence analysts, so tools that provide focus are useful. One way to determine focus is by computing significance. In the context of a known model, new data can be placed on a spectrum defined by: normal, anomalous, interesting, novel, or random; and significance is greatest towards the middle of this spectrum. However, significance also depends on the mental state of the analyst (and the organization). A framework for understanding significance is defined, and its impact on the knowledge discovery process is explored.