Unstructured Data

Unstructured data can be challenging to extract meaningful insights from for two reasons. First, such documents lack a predictable structure meaning that relevant information may reside at any location. Second, relevant information invariably co-exists with irrelevant information.

 

To address these difficulties we apply a two-phase approach involving a pre-processing pipeline (to filter out irrelevant information) followed by the deployment of rules-based and/or artificial intelligence algorithms (to extract relevant information).