Information Extraction Module

The Information Extraction Module can now parse any natural language input text and identify the relationship types that exist between the various pairs of entities identified in the text. This is achieved using the new Relationship Extractor stage.

Relationship Extractor can identify relationships between these entity types:

• Person
• Organization
• Location

The Information Extraction Module now has an Administration Utility command that evaluates and trains new and existing models. The command is iemodel trainAndevaluate model. It replaces the iemodel evaluate train_model command which evaluated only an existing model.

The Information Extraction Module now has separate stages for text categorization and entity extraction. Previously, both these functions were performed in the Information Extractor stage. Now there are two stages: Entity Extractor and Text Categorizer. Use Entity Extractor to extract entities, such as names and addresses, from strings of unstructured data. Use Text Categorizer to assign custom categories to unstructured content or plain text.