1. Administration and Configuration Guide
  2. Use Cases
  3. Customer Risk Scoring
  4. Obtaining SAR Labels for Customer Risk Scoring

5.6.4 Obtaining SAR Labels for Customer Risk Scoring

Obtain the SAR Information for Sandbox

Disposition/SAR information of the historical alerts that are required in sandbox for the purpose of supervised machine learning model training. SAR information acts as a target/depended variable for the model training.

Populate Investigated Entity Details

Obtain the SAR from CRR/ECM

Use aif.load_sar_data () API to load the Suspicious Activity Report (SAR) entities details from the Compliance Regulatory Reporting (CRR) application and Non-SAR entities from ECM into Compliance Studio.

Note:

“aif” is just a package that is available as part of compliance Studio.
The data will be loaded into the aif_investigated_entity table.

Figure 5-74 Aif Load SAR Data


Description of Figure 5-74 follows
Description of "Figure 5-74 Aif Load SAR Data"

The following parameters are the input value for the paragraph:
  • from_date: From date range in YYYYMMDD format for SAR/Alert creation date.
  • to_date: To date range in YYYYMMDD format for SAR/Alert creation date.
  • CRR_conn: CRR Connection object.
  • ECM_conn: ECM Connection object.

    Note:

    • Register Oracle wallet entries/aliases for CRR & ECM Atomic schema to connect within Compliance Studio
    • Use the aliases mentioned here to create/register entries. If aliases are being created with some other name, use them accordingly in the Admin Notebook.

Obtain the SAR from the CSV file

Use aif.load_sars_from_csv() API to load the SAR and Non-SAR entities into a CSV file.

Figure 5-75 Aif Load Sars from CSV


Description of Figure 5-75 follows
Description of "Figure 5-75 Aif Load Sars from CSV"

The following parameters are the input value for the paragraph:
  • filename: Complete path of the CSV file.
  • headerIncluded: This parameter has two options: Y or N. If the file has data with the header, then Y or N.

    Note:

    • The date should be in YYYYMMDD HH24:MI:SS format.
    • Records should be comma-separated (CSV).
Ensure that the following columns are available in the CSV files with the required values:
  • ENTITY_ID: Customer Id or Account Id
  • SUSPICIOUS_FLAG: This parameter has two options: Y or N. If E-file for Regulatory body has been sent for Customer or Account, then Y or N.
  • ALERT_DATE: SAR/EVENT generated to date from Customers and Accounts
  • CREATED_ON: CSV file creation date
  • CREATED_BY: CSV file created by
  • UPDATED_ON: CSV file updated date
  • UPDATED_BY: CSV file updated by
  • LABELLED_SCENARIO: This value has the following options:
    • CUST: For customer-level SAR
    • ACCT: For account level SAR
  • ENTITY_CD: This value has the following options:
    • If entity type is customer
    • If entity type is the account

Obtain the SAR classification from the CRR database

The aif.get_case_data_and_sar_classification() API gets SAR classification from CRR schema, merge with entity ID (Customer ID) in ECM, and stores as metadata in Compliance Studio schema table, aif_case_information.

Figure 5-76 Aif Get Case Data


Description of Figure 5-76 follows
Description of "Figure 5-76 Aif Get Case Data"

The aif_case_information table columns are as follows:
  • ENTITY_ID
  • CASE_ID
  • SAR_CLASSIFICATION
  • FILING_AM
  • CONTINUING_SAR
  • FILING_DATE
The following parameters are the input value for the paragraph:
  • from_date: From date range in YYYYMMDD format.
  • to_date: To date range in YYYYMMDD format.
  • CRR_conn: CRR Connection object.
  • ECM_conn: ECM Connection object.
  • AIF_conn: AIF Connection object.

Format: cx_Oracle.connect(<db_user/db_password@tns>)

On successful execution of the paragraph, the details will be loaded in the aif_case_information table.

Note:

  • Register Oracle wallet entries/aliases for CRR and ECM Atomic schema to connect within Compliance Studio.
  • Use the aliases mentioned here to create/register entries. If aliases are being created with some other name, use them accordingly in the Admin Notebook.