Golden-Proxy Dataset
Jump to navigation
Jump to search
A Golden-Proxy Dataset is a test dataset that is a hand-crafted proxy dataset specifically designed to target known failure modes, edge cases, and critical test scenarios.
- AKA: Golden-Proxy Data, Golden Test Set, Golden Test Dataset, Hand-Crafted Test Dataset, Reference Proxy Dataset, Gold Standard Test Dataset.
- Context:
- It can typically target Known Failure Modes identified through production analysis.
- It can typically provide Deterministic Test Results for quality assurance tasks.
- It can often include Expert-Designed Scenarios for comprehensive testing.
- It can often validate Critical System Behavior under controlled conditions.
- It can often complement Golden-Organic Datasets in evaluation frameworks.
- It can often support Automated Testing Pipelines with expected outcomes.
- It can range from being a Simple Golden-Proxy Dataset to being a Complex Golden-Proxy Dataset, depending on its scenario sophistication.
- It can range from being a Functional Golden-Proxy Dataset to being a Non-Functional Golden-Proxy Dataset, depending on its test focus.
- It can range from being a Positive Golden-Proxy Dataset to being a Negative Golden-Proxy Dataset, depending on its expected behavior.
- It can range from being a Unit-Level Golden-Proxy Dataset to being a System-Level Golden-Proxy Dataset, depending on its test scope.
- ...
- Examples:
- Edge-Case Test Suites, such as:
- Failure Mode Test Datas, such as:
- Compliance Test Datasets, such as:
- ...
- Counter-Examples:
- Golden-Organic Dataset, which comes from production rather than hand-crafting.
- Random Test Dataset, which lacks targeted design.
- Generated Synthetic Dataset, which is algorithmically created rather than hand-crafted.
- See: Proxy Dataset, Benchmark Dataset, Test Dataset, Adversarial Proxy Dataset, Golden-Organic Dataset, Edge-Case Generator, Evaluation Data Curation Task, Hand-Crafted Dataset.