Difference between revisions of "Data Sourcing"
From Open Risk Manual
Wiki admin (talk | contribs) |
Wiki admin (talk | contribs) |
||
Line 7: | Line 7: | ||
* Lack of [[Risk Data Standards]] can substantially complicate the sourcing of relevant data | * Lack of [[Risk Data Standards]] can substantially complicate the sourcing of relevant data | ||
* Incomplete or otherwise faulty data sourcing processes can [[Model Bias | potentially bias]] the data samples and thereby limit model quality. | * Incomplete or otherwise faulty data sourcing processes can [[Model Bias | potentially bias]] the data samples and thereby limit model quality. | ||
+ | * [[Data Provenance]] | ||
== References == | == References == |
Latest revision as of 11:28, 5 January 2022
Definition
Data Sourcing (also Data Collection) is the process of extracting data from external or internal (front/back office systems) comprising an institution's Data Infrastructure for diverse purposes of Risk Management, Portfolio Management and/or other business objectives.
Data Sourcing is particularly important for the purposes of Model Development and Model Validation.
Issues and Challenges
- Lack of Risk Data Standards can substantially complicate the sourcing of relevant data
- Incomplete or otherwise faulty data sourcing processes can potentially bias the data samples and thereby limit model quality.
- Data Provenance