Also known as the Computable News Project.

It is a maintained hypothesis that unstructured data (for instance, text) has a major impact on investment decision making. The purpose of this project is to structure unstructured information and to use that structure to aid investment decisions. Named entity recognition and text summarisation are key techniques being used in the creation of the relevant software.

The project has a number of separate uses:

• One use is to assist investors discriminate between value relevant and non value relevant information. 
• A second use is to reduce false positive alerts in a surveillance process by helping surveillance analysts (and the algorithms with which they work) to explain price movements.
• A third use is to quickly sort the chaff from the wheat in a market that in the midst of continuous disclosure regimes encourage some parties to over report which makes it more difficult to recognise valuable information when it is made available.

We are currently in the process of signing up an industry research partner.