PEP leverages the Analytics Grid to quantify and qualify real-world events with high precision and speed. The detection of "unusual" Topic patterns in real time is
typically the first stage in a highly extensible process that employs domain-specific rules and a powerful model for extracting facts and figures for reasoning. PEP rules enable outputs to be generated that incorporate
real time spatial and temporal information about Topic Models using typical first-order logic functions.
Rules Engine
Psydex leverages commercially available Rules Engines and also partners with CEP (Complex Event Processing) providers .
Rules are organized around Domain-specific Event Types and involve relative and temporal proximity of Topics across space and time, statistical thresholds, association with groups and other topics, and conditional criteria.
Satisifed rules result in the extraction of relevant attributes and values that drive downstream processing and decision making.
Attribute Extraction
Whether you need to identify the companies involved in an acquisition, or the location and magnitude of the latest earthquake, Psydex extracts facts and figures with high precision in real time.
When an event is detected, a set of event-specific attribute extraction functions are applied.
Attribute Extraction uses best of breed techniques and technologies, including Regular Expressions, Natural Language Processing (NLP), Entity and Relationship Extraction to identify and extract attributes that define the event.
Psydex supports multiple extraction functions for each Event Attribute to achieve the highest precision based on the source and format of the originating content.
Psydex Corporation - One Atlantic Center - 1201 West Peachtree Street, Suite 3000 - Atlanta, Georgia 30309-3455