Skip to main content

Creating a Concept Parser Project - Initial Setup

The prerequisites for creating a Concept Parser are the following:

  1. A Classifier or category determiner on which the SOC Project would've run (or at-least a Classifier Concept with values).

  2. Datasets for training and actual Project Data based on which the Concept shall be Parsed.

So, let us look at a typical Concept Parser project creation flow:

From the ‘All Project’ listing screen, the user needs to start the flow by clicking on the ‘Create New Data Classification Project’ button.

Step 1: Provide the Name & (Optionally) the Description

In the first screen, as in the case for Semantic Object Classification Projects, the user needs to provide the name and optional description of the Project. The user must also choose the correct Project Type and press the ‘Next Step’ button.

Step 2: Add Users and specify their Entitlements

This step is also the same as in Semantic Object Classification Project flow. The logged-on user will be treated as the default Admin and can add Reviewer(s) and Approver(s) as per the Four Eyes check. Once the entitlements have been added, press the ‘Next Step’ button.

Step 3: Select the Concept to be Parsed

In this step, the user must first select a Catalog, then- one child Semantic Object and finally, one Concept from the list of Concepts which is to be parsed into Sub-Concepts. Then the ‘Next Step’ is to be pressed once the selection is made.

Step 4: Select the Sub-Concept Classifier, it’s Source and Column

Now, the user must select the Classifier (remember the prerequisites we’ve mentioned in Introduction) in the Catalog and its Data Set source. So for this step to be completed, the user needs to complete these small sub-steps / tasks:

  1. Select the Classifier type Concept.

Note that you will only see a value here IF there is a Concept in the Semantic Object that you selected earlier which is a ‘Classifier,’ which essentially means that SOC project should’ve run successfully on it to provide relevant values to that Classifier. It is based on this Classifier that the Sub-Concepts would vary, whose values are to be predicted eventually in this project.

2. Then, select a single Data Set which contains the reference values for this Sub-Concept.

3. Once the Data Set is decided, the user must select a single column, which contains the values of the Sub-Concept.

Selecting a column here will show the profiler information on the right so you can re-verify if it is indeed the correct data column.

On pressing ‘Next Step,’ a Progress box will appear. In the background, it will load the values from the selected column. The loading may take a little time due the job queueing process.