Auto Classification for MDM 360 Applications > Introduction to Auto Classification for MDM 360 Applications recipe
Introduction to Auto Classification for MDM 360 Applications recipe
The Auto Classification for MDM 360 Applications recipe is based on REST and SOAP APIs. Use the recipe to automatically classify records from Informatica Master Data Management (MDM) applications using a Large Language Model (LLM).
The recipe uses Pinecone VectorDB and OpenAI to identify classifications and categories that are semantically similar to your records. It is designed for use with the MDM validation and enrichment framework, employing Objective Groups created in Business 360 Console.
The recipe functions effectively for both standard relationships and hierarchies, offering capabilities to either auto-suggest relationships or auto-assign the most similar ones.
For example, the recipe automatically identifies and assigns categories to items in Product 360. In this context, you can utilize standard taxonomies, such as UNSPSC, eClass, Google, or Amazon, as well as custom taxonomies, with the category business entity serving as the source.
The recipe adopts a two-step approach as follows:
1Indexing: Feeds a Pinecone vector database with the classifications.
2Searching: Matches business entity records with classifications by first searching the Vector database for candidates, retrieving semantically similar records from VectorDB, and then using OpenAI to refine the candidate list for precise similarity scoring.
The following image illustrates the architecture diagram of Auto Classification:
