Node Description

Correlation Repair Node

The Correlation Repair node replaces missing Product Attributes with the Attributes found in highly correlated Products.

The Correlation Repair node compares Products that are missing the selected Attribute (Missing Attribute Products) from all of the other Products in the Input Product Array having a value for the selected Attribute. For example, if the user wishes to repair the ‘Category’ Attribute column, then all Products with missing Category values are compared against the remaining Products having a Category value.

This Premium Node is not available as part of the free Community Edition. Premium Nodes help clean and connect real-world data to Market Simulations, and provide advanced Market Science analysis. Note that these descriptions are often deliberately vague.

Downloads

Workflow Image

Correlation Repair

The Correlation Repair node replaces missing Product Attributes with the Attributes found in highly correlated Products. For example, if the user wishes to repair the ‘Brand’ column, then all Products with missing Brand values are compared against similar Products having a Brand value.

Inputs

Input Product Array

The complete set of all Products containing both Products that are missing Attribute values (Missing Attribute Products) as well as Products that are not missing Attribute values.

Input Product Similarity Rankings

The list of all Product-to-Product rankings. That is, from the Customers who looked at Product01 how did they rank Product02. Alternatively, from the Customers who looked at Product01 what is the probability that they also looked at Product02 (sorted to provide a rank order).

Node

Configuration

The user specifies which column from the Input Product Array contains the Missing Values that needs to be repaired.

Outputs

Output Product Array

The Output Product Array corresponds to the Input Product Array but is updated to include the repaired Missing Attribute.

Output Product Similarity Rankings

The list of all Product-to-Product rankings. The Output Product Similarity Rankings is unchanged and identical to the Input Product Similarity Rankings. This table is merely passed-through as a convenience so that several Correlation Repair nodes can be chained together – each designed to repair a different Attribute column.