Data quality rules that can be used specifically to classify objects based on their individual input values.

Classify_Product

General description

Classifies a product to a structure group based on its "Long Description". A threshold (default value of 70%) defines the score needed to decide whether the classification should happen or not.

NOTE: This is a DEMO rule and classifies based on the "Sample Data Consumer Electronics" demo data. For own implementations, create a new classifier and change the configuration of the classifier transform (cls_get_category) to point to your classification model.

NLP Classifiers need to be trained with actual live data. This happens by mapping e.g. long descriptions of products to structure groups of Product 360. The broader the variety of long descriptions that have been mapped to a specific structure group value the better is the classification of new items working.

Input ports

Long_Description	Long description that will be used to classify the object against a certain structure group.

Output ports

StructureGroup_ID	Returns the identifier of the structure group found for the product.
StructureGroup_Name	Returns the English name of the structure group found for the product.
StructureGroup_Score	Returns the probability score for the classification to the structure group found. (QualityStatusEntry.Score)
Out_Status_Code	Returns the overall Status Code after the rule execution (OK or Failed). (QualityStatusEntry.Status)
Out_Status_Message	Returns the overall Status Message after the rule execution. (QualityStatusEntry.Message)

Meta data and reference tables

cs_classify_products_demo	Sample classifier based on the Consumer Electronics "Sample Data Consumer Electronics" demo data.
Error_messages_by_Language	Contains a 4 digit error code plus a language to indicate the preferred error message to be output.

Example usage

Check an item for a proposed structure group of the "Sample Data Consumer Electronics" structure system.

The "English Long Description" (Long_Description) of the item is:

"46 inch multimedia television set (Full-HD) with high-quality aluminum front frame and contrast pane - comes equipped with Dual HDTV multi-tuner and digital video recorder (integrated hard drive)"

Since the classifier of the rule has been trained with multiple possible long descriptions for all the structure groups of the system it suggests that the item should be assigned to the structure group called "HDTV Series"

Example output

StructureGroup_ID	1267688379455
StructureGroup_Name	HDTV Series
StructureGroup_Score	1.0000
Status_Code	OK
Status_Message	No Error

Identify_Language