Big Data Quants - Is News Analytics Another Form of Riding the Price Waves?

It was the cover story of the March-13 issue of the Wilmott magazine.  In short, it is about the extraction of insight from unstructured data - news, sentiment, ..

Many people may think news analytics is for event-driven--black-box trading, but Reuters, for example, has something different in mind. Something that fits to white-box quant finance, IMO. News-based factor generation for intelligent decisions? A technology to support multi-factor strategies.

I started thinking about this with Riding the Price Waves (process more real market information for better parameter identification) and in my most recent post Should Quants Learn More About Machine Learning I introduced the metaphor of context that "closes the gap" between the quantitative regime and the market dynamics. With big data as one of the contextual technologies.

Simplified, one could imagine the transformation of unstructured information (about the macro-economic dynamic, regulatory decisions, social network information  ...) into a decision tree. Walking through the tree you get to the best quantitative regime and start the quantitive workflow. Unfortunately it is not that simple, because of uninformative data and the difficulty to decide which model does explain the data set best.
And clearly, news analytics need to deal with many entities, relevance, events, sentiment, ...

But, it is my strong belief, news analytics should not strive for extracting, say, trading decisions directly and automatically ... but provide higher information to empower quants and risk professionals to use better model instances. This may also constrain the parameter space to a realistic size?

In a bottom up approach we start with the core workflow in practical quantitative finance simulations as follows (chapter 15 of a Workout in Computational Finance)

  1. choose the right model for the movement of the underlying 
  2. determine its parameters (the better the market data, the better the valuations)
  3. use the right numerical scheme 
  4. Modify the market data for scenario analysis, stress tests for trading and risk management decision support
Information about an expected market behavior will obviously support 2. and 4. (but also 1. automated model select seems possible).

Summarizing, I see news analytics as an important factor in a muti-factor strategy with white-box models in quantitative approaches. Maybe more important than as part of the armory of the event-driven-blackbox trading community.

UnRisk has organized instruments-models-methods orthogonally and structures nested groups of instruments and nested groups of scenarios. It also applies robust schemes for inverting. Welcome, contextual technology. Welcome, news analytics.