Machine Learning Product Recommender
Overview
The Machine Learning Product Recommender compares customer purchase history or other preference attributes to a set of product attributes, and generates one or more product recommendations for each customer. The tool has two inputs: customer data, and product data. Each input should contain a unique ID field, and one or more attribute fields. Customer preferences may be represented as product IDs or as keywords. Product attributes are the keywords associated with each product ID.
The tool outputs a specified number of customer ID/product ID pairs for each unique customer ID, along with rank and match scores for each pair.
Configuration parameters
The Machine Learning Product Recommender has two sets of configuration parameters in addition to the standard execution options.
Customer Data
Parameter | Description |
|---|---|
Number recommendations | The number of recommendations to generate for each customer. |
Customer ID field | Field containing the unique ID for each customer record. |
Customer history | One or more fields containing customer preference attributes, expressed as product IDs or offer IDs. Each field can be assigned a Weight, with higher weights given high priority in matching. |
Customer preferences | One or more fields containing customer preference attributes, expressed as keywords. Each field can be assigned a Weight, with higher weights given high priority in matching. |
Product Data
Parameter | Description |
|---|---|
Product ID/Offer field | Field containing the unique ID for each product or offer. |
Product ID/Offer Rank field | Optional field containing the ranking for each product or offer. |
Keyword/attribute fields | One or more fields containing product attributes, expressed as keywords. Each field can be assigned a Weight, with higher weights given high priority in matching. |
Configure the Machine Learning Product Recommender
Place a Machine Learning Product Recommender tool on the canvas, and then attach the customer and product data inputs to the C and P input connectors.
Select the Machine Learning Product Recommender tool.
Go to the Customer data tab.
Specify the desired Number of recommendations, and then select the Customer ID field.
Select the Customer history grid and Customer preferences grid, and choose fields containing customer product history and preference attributes. You may optionally specify a Weight for each field.
Select the Product/Offer data tab, and then choose Product/Offer ID field and optionally a Product/Offer Rank field.
Specify Keyword/Attribute fields. You may optionally specify a Weight for each product attribute.
Optionally, go to the Execution tab, and then set Web service options.