Supervised Scoring

5.6.7.2 Supervised Scoring

This is a pre-seeded batch and will be available in all workspaces ( production & sandboxes )
This Batch is to be executed in the Production workspace.

Batch and Task Parameters

The batch contains the following tasks:

Task 1: Scoring_Data_Load
Task 2: ML_Scoring
Task 3: ECM_Event_Processing

Task 1: Scoring_Data_Load, Task Parameters

Objective folder for this task:

Home / Modeling / Pipelines / AIF Batch Framework / Supervised ML /
Scoring Data

Optional Parameters:
- from_date: Start date for Scoring Data lookup in DD-Mon-YYYY format.
- to_date: End Date for Scoring/New Data lookup in DD-Mon-YYYY format.
Example: from_date=01-Jul-2020,to_date=31-Jul-2021
Optional Parameters can be edited from the Schedule Batch option.
Change any other batch /task parameters, except Optional Parameters.

Figure 5-93 Edit Task for Scoring Data Load

Description of "Figure 5-93 Edit Task for Scoring Data Load"

Task 2: ML Scoring, Task Parameters

Objective folder for this task:
```
Home / Modeling / Pipelines / AIF Supervised ML / AIF
```
- Navigate to respective model group/scenario folders for actual model templates.
- Optional Parameters:
  - osot_end_month: Specify the scoring data month in YYYYMM format. If not specified by default latest month data available in the table will be picked up for scoring.
  - threshold: Input threshold or cutoff to create events. Events will be created if the score of an entity exceeds the threshold. Example: 0.7
  - from_date: Start date for Scoring Data lookup in YYYYMM format.
  - to_date: End Date for Scoring/New Data lookup in YYYYMM format. Example : from_date=202007,to_date=202007
- Optional Parameters can be edited from the Schedule Batch option.
- Choose Link Types as Scoring
- Do not change any batch/task parameters, except Optional Parameters.
  
  Figure 5-94 Edit Task for ML Scoring
  
  Description of "Figure 5-94 Edit Task for ML Scoring"

Task 3: ECM_Event_Processing, Task Parameters

Objective folder for this task:

Home / Modeling / Pipelines / AIF Batch Framework / Supervised ML /
Event Processing

This task does not take any optional parameters.
Do not change any other batch/task parameters.

Figure 5-95 Edit Task for ECM Event Processing

Description of "Figure 5-95 Edit Task for ECM Event Processing"
After scoring for supervised customer risk scoring, the outputs are stored in the AIF_ENTITY_SCORE table.
Alerts generated above thresholds are moved to the following tables for case management integration:
- FCC_AM_EVENTS
- FCC_AM_EVENT_DETAILS
- FCC_AM_EVENT_ENTITY_MAP
- FCC_AM_EVENT_BINDING

Cleanup Steps in case of running the Scoring Process twice

In case the user wants to run the Scoring Process for the same FIC_MIS_DATE and same MODEL_GROUP_NAME twice, the following cleanup steps should be performed first:

Remove the existing events:

delete from fcc_am_event_binding where v_event_cd in (select v_event_cd
from fcc_am_events where prcsng_dt='DD-Mon-YYYY');
delete from fcc_am_event_entity_map where v_event_cd in (select
v_event_cd from fcc_am_events where prcsng_dt='DD-Mon-YYYY');
delete from fcc_am_event_details where n_event_cd in (select v_event_cd
from fcc_am_events where prcsng_dt='DD-Mon-YYYY');
delete from fcc_am_events where prcsng_dt='DD-Mon-YYYY');

Get the child tables which contain scoring results:

select D_FIC_MIS_DATE, V_MODEL_GROUP, V_OUTPUT_TABLE_NAME,
V_OUTPUT_TABLE_NAME_ALL_ENTITY
from aif_entity_score
where d_fic_mis_date ='DD-Mon-YYYY'
and model_group_name='<Model_Group_Name>';

Drop all child tables manually listed in V_OUTPUT_TABLE_NAME and V_OUTPUT_TABLE_NAME_ALL_ENTITY columns from the result of the above query :
```
drop <Child_Table_Name>;
```
Delete the parent entry from aif_entity_score:
```
delete from aif_entity_score where d_fic_mis_date='DD-Mon-YYYY'
```