Search Results ar_cmgt_score_weights

Overview

The AR_CMGT_SCORE_WEIGHTS table is a Receivables (AR) module database object that stores the weighting factors applied to individual data points within a Collections Management scoring model. In Oracle EBS Release 12.1.1 and 12.2.2, scoring models are used by the Collections Management (Advanced Collections) functionality to compute predictive or rule-based scores that rank delinquent customers and drive the assignment of collection strategies. Each row in this table represents the relative importance assigned to a single data point when the model aggregates its score.

The table resides in the AR schema and is documented as VALID in ETRM 12.2.2. The heuristic Data Vault classification mined from its foreign key structure is standalone; from a modeling perspective, this object is best treated as a standalone hub-style or reference structure rather than a dependent satellite, because the mined relationship data does not identify parent keys outside the table itself. The composite primary key implies that weight values are entities defined by the combination of a scoring model and a data point, not by either element alone.

Key Information Stored

The documented physical schema lists eight columns. The columns of greatest functional significance are:

• SCORE_MODEL_ID — Identifies the Collections scoring model to which the weighting belongs. This is the leading component of the primary key and ties the weight to a specific model definition.
• DATA_POINT_ID — Identifies the individual data point (the customer, account, transaction, or behavior attribute) being weighted within the model. This is the second component of the primary key.
• WEIGHT — The numeric weighting factor applied to the data point when the model computes its score. This is the substantive business value of the row and the target of most queries against the table.
• LAST_UPDATE_DATE — The date the weight row was most recently modified; useful for audit and change tracking.
• LAST_UPDATED_BY — The user who performed the most recent update.
• LAST_UPDATE_LOGIN — The login session associated with the last update.
• CREATION_DATE — The date the weight row was created.
• CREATED_BY — The user who created the row.

The surrogate primary key is AR_CGMT_SCORE_WEIGHTS_PK, defined on the composite of SCORE_MODEL_ID and DATA_POINT_ID. A unique index, AR_CMGT_SCORE_WEIGHTS_U1, is documented on the same two columns, indicating that this pair is simultaneously the business-key candidate and the primary key. There is no separate single-column surrogate identifier; the model-to-data-point combination is the natural identifier of a weight.

Common Use Cases and Queries

Typical usage centers on reviewing, comparing, and auditing model configuration. Administrators and analysts frequently need to see which data points carry the most influence in a model, or to compare weights across two models to understand why their scoring outputs differ.

A common pattern retrieves all weights for a given model, ordered by importance:

• SELECT data_point_id, weight FROM ar.ar_cmgt_score_weights WHERE score_model_id = :model_id ORDER BY weight DESC;

A second pattern looks up the weight assigned to a specific data point within a model:

• SELECT weight FROM ar.ar_cmgt_score_weights WHERE score_model_id = :model_id AND data_point_id = :point_id;

Reporting use cases include validation that all required data points for a model have a weight row, detection of zero or negative weights that may indicate misconfiguration, and change-history reporting driven by LAST_UPDATE_DATE, LAST_UPDATED_BY, and CREATION_DATE. Because the table is described as storing weights "for each data point in a scoring model," completeness checks — joining model data points against this table to find missing weights — are a frequent maintenance exercise.

Related Objects

The metadata classifies this table as standalone, so no parent foreign keys were mined. Its practical dependencies are driven by the two key components rather than by enforced referential constraints. The most significant related objects are:

• AR_CMGT_SCORE_MODELS (or the corresponding scoring model definition object) — joined on SCORE_MODEL_ID to resolve the model name and attributes for a weight row.
• AR_CMGT_DATA_POINTS (or the corresponding data point definition object) — joined on DATA_POINT_ID to resolve the human-readable name and source of the weighted attribute.
• AR_CMGT_SCORE_WEIGHTS_U1 — the unique index enforcing uniqueness of the SCORE_MODEL_ID / DATA_POINT_ID business key.
• AR_CMGT_SCORE_WEIGHTS_PK — the primary key constraint on the same columns.
• Advanced Collections scoring engine / API layer — the runtime consumer that reads WEIGHT values when computing scores for customers and strategies.

Direct SQL against this table should be read-oriented; inserts and updates are normally performed through the Collections Management setup forms or the supported scoring model APIs to preserve model integrity.