The world of unstructured data appears to hold out new possibilities for acting on future events.
If by unstructured we suggest data that is not already conventionally indexical (and in a sense of course all data is already structured), then data such as social media and other web-based forms is increasingly text mined for emerging phenomena. So, for example, commercial uses of text mining of social media data would look for associations and links with ‘churn’ to identify those with a propensity to shift to a new provider.
In the security context, of course, this technique that is all but ubiquitous in commercial spheres has become a matter of political an public debate and concern.
Is the use of software platforms that make sense of large volumes of unstructured data a different matter when we talk about its use for security purposes? And if it is then how do we disentangle it from the mesh of data analytics that now makes up the very fabric of our daily lives?
Volha and I spent time in London learning more about the use of predictive analytics, decision trees, partitioning and noise detection.