Description
Check feature availability and target ordering at prediction time.
Model validation depends on when information became available and how records or entities enter splits. A timestamp inequality or group overlap count can identify one local issue without proving absence of every form of leakage.
Check feature availability and target ordering at prediction time.
When to use this tool
Use Target Leakage Explicit Time Order when checking an explicit feature-time or split-membership condition.
How to use the tool and how it works
- Enter featureAvailable, predictionTime, targetObserved, allowSimultaneous. Use the units and conventions below.
- Run the tool. Correct any format, range, or cross-field input error before interpreting the output.
- Read featureEligible, targetInFuture. Inspect featureEligible, targetInFuture before interpreting the associated values.
Check feature availability and target ordering at prediction time. featureEligible uses <= when allowSimultaneous, otherwise <; targetInFuture always requires targetObserved > predictionTime.1
Preprocessing and feature construction must respect training and evaluation separation. A local time inequality or leave-one-out term is only one part of a leakage review.2
| Input field | Meaning and units | Accepted input |
|---|---|---|
| featureAvailable | Feature Available | Required. Finite number. |
| predictionTime | Prediction Time | Required. Finite number. |
| targetObserved | Target Observed | Required. Finite number. |
| allowSimultaneous | Allow Simultaneous | Required. Boolean. |
| Output field | Interpretation | Returned shape |
|---|---|---|
| featureEligible | Feature Eligible | Required. Boolean. |
| targetInFuture | Target In Future | Required. Boolean. |
{
"input": {
"featureAvailable": 9,
"predictionTime": 10,
"targetObserved": 12,
"allowSimultaneous": false
},
"output": {
"featureEligible": true,
"targetInFuture": true
}
}Limitations and assumptions
- Availability timestamps and simultaneity rules are supplied; no full leakage audit or causal validity is established.
- Inputs must satisfy both the field types and the cross-field conditions above. Invalid input causes an error; a successful false flag or documented placeholder is a different result.
- Numeric calculations use binary floating point unless exact decimal-integer arithmetic is explicitly stated. No hidden measurement tolerance or decimal rounding is added; nonfinite or out-of-contract results are rejected. Numeric signed-zero results are exported as zero.
Alternative or Complementary methods
Use a complete data-provenance and validation-design review, including label horizons, entity dependence and preprocessing fit scope.
References
-
Calculation and input contract: data-science/target-leakage-explicit-time-order — tool.io
-
Common pitfalls: data leakage — scikit-learn developers
Similar or alternative tools
- Target Observation Window Width
Return target observation-window width and eligible duration.
- Train Test Time Gap
Check a supplied train/test separation against an embargo.