Building of Text-Mining Models

• Based on pattern recognition and genetic algorithms, the Term Model Wizard enables users to quickly and easily build text-mining models for automated data extraction; no prior programming or text-mining experience is required.
  
  
• Praedea provides a set of financial data text-mining models for use in the extraction of financial data; users customize these models and build additional models to their requirements.
  
  
• Users auto-extract and validate data from an initial set of documents.
  
  
• Through the processing of additional documents the text-mining models are optimized to recognize additional data reporting patterns.
  
  
• As optimization and re-optimization occurs, the text-mining models are capable of delivering near 100% extraction accuracy for financial table data and other consistently-presented textual data.
  
  
• Text-mining models also apply “company-specific memory,” the ability to recall how specific companies report information, to further enhance accuracy.