Skip to main content

Introduction to Classification Projects

There are two types of Classification Projects:

  1. Semantic Object Classifier: Classify records in a Data Set using a ‘Classifier’ for a specific Semantic Object mapped to multiple Data Sets.

  2. Concept Parser: Parse or split information embedded inside Concepts (through its mapped columns) into multiple new Concepts.

We will cover the first one in this section which allows a user to run Classification on a Selected Semantic Object of a Catalog. Let’s understand this through an example:

  • In the case of a large Heavy Machinery / Electronics / Manufacturing company, assume there is a huge inventory list and the legacy description and details are non-standardized, haphazard, providing no clue about the Product type. Example:  – Is it a beaker, is it a Pipe, is it a Funnel, is it a Tripod etc. It is difficult for the company to identify  & group individual products. 

  • These subcategories of the main product are all Classes or categories to which the main Product (our Semantic Object) can be categorized(Classified), based on certain inputs that are other attributes of the Product. The Semantic Object Classification or SOC Project does this very thing.

  • Here’s another example: A medical company, which has to categorize patients into Low, Medium or High Risk category for a certain ailment, can do it on the basis of inputs from the Patient data (Our Semantic Object). This is again possible by running an SOC Project.

All this would take a long time, if done manually, and the process- prone to oversight and errors, but can be done easily through such a Project set-up.

Now let us cover the workflow of creating a Semantic Object Classifier, also known as the SOC Project.