This topic describes how to configure matching recall in the procedure of creating instances, configuring matching and filtering strategies, and deploying and testing models.

Prerequisites

AutoLearning is authorized to access Object Storage Service (OSS). For more information, see OSS authorization.

A Tablestore (OTS) instance is created. For more information, see Create instances.

Step 1: Create an instance

  1. Log on to the Machine Learning Platform for AI console.
  2. In the left-side navigation pane, choose AutoLearning > General Purpose Model Training.
  3. On the AutoLearning page, click Create Instance.
  4. On the Create Instance page, set the following parameters.
    Parameter Description
    Instance Type In the Instance type field, select Rec-Matching System. AutoLearning provides the following instance types:
    • Image Classification
    • Rec-Matching System
    Instance Name The name must be 1 to 20 characters in length and can contain letters, underscores (_), and digits. It must start with a letter.
    Example Description The description of the instance, which helps distinguish different instances.
    Storage Dependency To use matching recall, you must store your training data in OTS of Alibaba Cloud. OTS is a database similar to KVStore for Redis. For more information, see Create tables.
    Instance Association Associate with an OTS instance.
  5. Click Confirm.

Step 2: Configure matching strategies

  1. On the AutoLearning page, click Open in the Operation column.
  2. In the Matching strategy configuration wizard, configure a matching strategy.
    Section Parameter Description
    Collaborative Filtering Recall Strategy name The name must be 1 to 20 characters in length and can contain letters, digits, and underscores (_). It must start with a letter.
    User-Item table Select a User table for the matching strategy. Move the pointer over the question mark icon next to User-Item table to view the schema of the table.
    Item-Item table Select an Item table for the matching strategy. Move the pointer over the question mark icon next to Item-Item table to view the schema of the table.
    Matching number The number of recalls to be returned under the matching strategy. The number must be an integer equal to or greater than 1. If the number of returned recalls is smaller than the specified Matching number, the actual number of returned recalls prevails.
    Semantic Recall Strategy name The name must be 1 to 20 characters in length and can contain letters, digits, and underscores (_). It must start with a letter.
    User-Topic table Select a User table for the matching strategy. Move the pointer over the question mark icon next to User-Topic table to view the schema of the table.
    Topic-Item table Select an Item table for the matching strategy. Move the pointer over the question mark icon next to Topic-Item table to view the schema of the table.
    Matching number The number of recalls to be returned under the matching strategy. The number must be an integer equal to or greater than 1. If the number of returned recalls is smaller than the specified Matching number, the actual number of returned recalls prevails.
    Self-Definition Recall Strategy Strategy name The name must be 1 to 20 characters in length and can contain letters, digits, and underscores (_). It must start with a letter.
    Item table Select a table for the matching strategy. Move the pointer over the question mark icon next to Item table to view the schema of the table.
    Matching number The number of recalls to be returned under the matching strategy. The number must be an integer equal to or greater than 1. If the number of returned recalls is smaller than the specified Matching number, the actual number of returned recalls prevails.
    AutoLearning supports the following matching strategies. You can select one or more strategies based on your requirements.
    • Collaborative filtering recall: a classic matching strategy that generates matching results based on the correlation between customers and products.
    • Semantic recall: a matching strategy that is widely used in content recommendation based on what topics the audience are interested in.
    • Self-Definition recall: a matching strategy that generates user-item correlations based on actual requirements.
  3. Click Add to strategy list. Then, the configured strategy is added to the Added strategies list.
  4. Optional:Repeat the preceding steps to add more matching strategies.
  5. Click Next step.

Step 3: Configure filtering strategies

  1. In the Data filter strategy configuration wizard, set the following parameters.
    Section Parameter Description
    Self-Defined Filter Strategy (U-I Filter) Strategy name The name must be 1 to 20 characters in length and can contain letters, digits, and underscores (_). It must start with a letter.
    User-Item table Select a table for the filtering strategy. Move the pointer over the question mark icon next to User-Item table to view the schema of the table.
    Self-Defined Filter Strategy (I Strategy Filter) Strategy name The name must be 1 to 20 characters in length and can contain letters, digits, and underscores (_). It must start with a letter.
    Item table Select a table for the filtering strategy. Move the pointer over the question mark icon next to Item table to view the schema of the table.
    AutoLearning supports the following strategies. You can select one or more filtering strategies based on your requirements.
    • A filtering strategy based on user-item correlations. If a specified user is returned in the matching results, the items that correspond to the user are filtered out. The following figure shows the table schema.Data format
    • A filtering strategy based on items. If the ID of an item is returned in the matching results, the item is filtered out. The following figure shows the table schema.Table schema
  2. Click Add to strategy list. Then, the configured strategy is added to the Added strategies list.
  3. Click Deploy and test.

Step 4: Deploy and test

  1. The system deploys the model as a service based on the configured matching strategies and filtering strategies.
  2. In the Test Module section, specify UserID and Recommendation results parameters.1
  3. Click Send test request.
  4. In the Debug information section, view the recommendation list. If the list meets your requirements, click Deploy to EAS to deploy the model to Elastic Algorithm Service (EAS) as a RESTful API.