Search Results ams_dm_score_results




Overview
Explore in AI Chat

Overview

The AMS_DM_SCORE_RESULTS table is a core data structure within the Oracle E-Business Suite (EBS) Marketing (AMS) module, specifically for its data mining and predictive analytics capabilities. It functions as a container for the summarized, aggregated results generated when a predictive model is executed against a target population—a process known as scoring. This table does not store individual record-level scores but holds the statistical summary of the scoring run, such as the distribution of scores across deciles. Its role is to provide a high-level, analytical view of a model's performance and output, enabling marketers to analyze the effectiveness of their predictive campaigns and make data-driven decisions.

Key Information Stored

The table's primary purpose is to store aggregated scoring metrics linked to a specific scoring instance. The key columns, as indicated by its constraints, include the unique identifier SCORE_RESULT_ID (primary key) and the foreign key SCORE_ID, which ties the summary results back to the parent scoring instance recorded in the AMS_DM_SCORES_ALL_B table. A critical analytical column is DECILE, which, as part of a unique key constraint with SCORE_ID, stores the decile ranking (typically 1-10) into which the scored population is segmented. While the provided metadata does not list all columns, typical data in such a table would include aggregate counts, sums, or average scores for each decile, allowing for analysis like lift and gains charts.

Common Use Cases and Queries

The primary use case is post-scoring analysis to evaluate a predictive model's performance. Marketers query this table to understand the concentration of high-probability targets in the top deciles, which is essential for optimizing campaign targeting and resource allocation. A common analytical query would retrieve the decile breakdown for a specific scoring run to calculate lift. For example:

  • Identifying the top-performing segments: SELECT DECILE, RESPONDER_COUNT, TOTAL_COUNT FROM AMS_DM_SCORE_RESULTS WHERE SCORE_ID = :p_score_id ORDER BY DECILE;
  • Joining to the scores header for contextual reporting: SELECT s.SCORE_NAME, r.* FROM AMS_DM_SCORE_RESULTS r, AMS_DM_SCORES_ALL_B s WHERE r.SCORE_ID = s.SCORE_ID AND s.MODEL_ID = :p_model_id;

This data is foundational for generating standard marketing reports on campaign prediction accuracy and return on investment.

Related Objects

The table maintains a strict, documented dependency on the central scoring header table via a foreign key relationship. The key related objects are:

  • AMS_DM_SCORES_ALL_B: This is the primary parent table. The foreign key AMS_DM_SCORE_RESULTS.SCORE_ID references AMS_DM_SCORES_ALL_B, linking each set of summarized results to its specific scoring instance. This relationship is critical for tracing summary results back to the model, campaign, and population context of the original scoring job.

As a summary table, AMS_DM_SCORE_RESULTS is typically an endpoint for analytical queries that source their transactional context from AMS_DM_SCORES_ALL_B and related model definition tables.