Attensity Corporation has made a fundamental breakthrough in converting unstructured text into structured tables with 95% or better accuracy (precision + recall). Rather than using statistical and probabilistic algorithms to extrapolate representations of meaning from word content and proximity, Attensity's technology understands English language by using computational linguistics to parse sentences into fundamental linguistic elements, and then analyzes these elements using sophisticated algorithms. This approach results in the following benefits:
Unprecedented accuracy allowing the extraction of events and entities from within documents, not just categorization of documents themselves
Identification of an event's attributes, i.e. what participated in an event, and how
High raw text throughput of 5MB/minute on 1GHz Intel CPU
Ability to handle noisy text: misspellings, poor grammar and bad punctuation
Attensity has been awarded five patents for its technology, with an additional twenty patents pending. For a technology whitepaper, contact us.