DeLongTest<T>
Models & Types · AiDotNet.Evaluation.Statistics
DeLong’s test for comparing two ROC curves.
Implements: IStatisticalTest<T>
For Beginners
DeLong’s test compares the AUC of two classifiers:
- Tests if the difference in AUC is statistically significant
- Specifically designed for comparing ROC curves
- Non-parametric and doesn’t assume normal distribution
- Can handle correlated data (same test set for both models)
How It Works
When to use:
- Comparing two binary classifiers on the same dataset
- When you want to know if AUC improvement is significant
- Medical diagnostic test comparisons
Constructors
| Constructor | Summary |
|---|---|
DeLongTest() |
Properties
| Property | Summary |
|---|---|
bool IsNonParametric { get; } |
|
bool IsPaired { get; } |
|
string Description { get; } |
|
string Name { get; } |
Methods
| Method | Summary |
|---|---|
StatisticalTestResult<T> Test(T[] predictions1, T[] predictions2, T[] actuals) |
Performs DeLong’s test comparing two sets of predicted probabilities. |