DataRefiner Factory, an artificial intelligence system, comprises several subsystems, various lexicons, and knowledge bases:

• Data Import Management: Import, map, and aggregate data from any source or database(Import Manager)
  
• Taxonomy Management: Build, or adapt, and maintain taxonomies, including categories, category hierarchy, attributes and values (Taxonomy Manager)
  
• Formalization: Automatically tokenize, translate, and standardize the product descriptions into a unified, normalized structure (Formalizer)
  
• Clustering: Automatically cluster the products into distinct clusters based on the content of the description (Clusterer)
  
• Classification: Automatically classify data according to any given taxonomy system (Classifier)
  
• Extraction: Automatically extract values from the product description into the relevant attributes according to a given taxonomy (Extractor)
  
• Enrichment: Automatically enrich selected attributes (Enricher)
  
• Deduping: Automatically identify products that share certain, pre-defined technical similarities by various user-defined strategies (Deduper)
  
• Normalized Description Generation: Automatically generates standardized descriptions that are systematic, informative, and pre-defined by category; descriptions can be created in any language, length, or structure (Descriptor)
• Data Export Manager: Export the rationalized product data to the target system in the desired structure, format, and communication protocol (Export Manager)

DataRefiner uses its high-performance product data quality solution to clean up dirty product data. For example, the product descriptions below are from a global company’s purchasing departments located in Korea, Israel, the United States, and Germany, actually describe the same product, an Allen screw.

Before DataRefiner: Messy raw data

1. DIN 912 10x1x30-2.9 mat304
2. ALLEN SCR M10x30 stainless steel
3. SOCKET BOLT M10x1 LG30 SS
4. M10x1mmx30mm SHCS-SS

After DataRefiner: Organized, classified, standardized, and rich product attributes

Category	Type	Screw Size	Thread Pitch	Head Style	Fastener Length	Drive System	Material
Screw	Machine Screw	M10	1mm	Cylindrical	30mm	Female Hex	Stainless Steel

Note: The attribute values are consolidated into one table assuming they meet the accepted tolerance.

Key Features Automatic — AI-based process ensures high accuracy and reliability Secure — company data remains behind firewall at all times Efficient — enables accurate data processing with minimal human intervention Effective — reuses acquired knowledge for ongoing data protection Precise — decisive product identification eliminates duplication and cuts procurement costs Compatible — operates with company's existing IT infrastructure Non-intrusive — simple integration process with no change in operational procedures Up-to-date — enables improved sales forecasting, planning, and product-demand tracking Universal — works with any language

Contact an InQuera representative to find out how DataRefiner can improve your product data quality.