Class TestTimeAugmentationResult<TOutput>

Namespace

AiDotNet.Augmentation

Assembly

AiDotNet.dll

Contains the result of a Test-Time Augmentation prediction, including individual and aggregated predictions.

public class TestTimeAugmentationResult<TOutput>

Type Parameters

TOutput

The type of prediction output (e.g., Vector for class probabilities).

Inheritance

object

TestTimeAugmentationResult<TOutput>

Inherited Members

object.Equals(object)

object.Equals(object, object)

object.GetHashCode()

object.GetType()

object.MemberwiseClone()

object.ReferenceEquals(object, object)

object.ToString()

Remarks

For Beginners: This class gives you both: 1. The final combined prediction (what you usually want) 2. All the individual predictions (for debugging or analysis)

You also get uncertainty information - if all 5 predictions were similar, you can be confident in the result. If they varied wildly, you might want to be more cautious.

Constructors

TestTimeAugmentationResult(TOutput, IReadOnlyList<TOutput>, double?, double?)

Creates a new Test-Time Augmentation result.

public TestTimeAugmentationResult(TOutput aggregated, IReadOnlyList<TOutput> individual, double? confidence = null, double? standardDeviation = null)

Parameters

aggregated TOutput

The combined prediction.

individual IReadOnlyList<TOutput>

All individual predictions.

confidence double?

Optional confidence score.

standardDeviation double?

Optional standard deviation of predictions.

Properties

AggregatedPrediction

Gets the final combined prediction after aggregating all augmented predictions.

public TOutput AggregatedPrediction { get; }

Property Value

TOutput

Remarks

For Beginners: This is the main result you'll use. It's the combination of all individual predictions based on the aggregation method (mean, median, vote, etc.).

Confidence

Gets the confidence score of the aggregated prediction (if available).

public double? Confidence { get; }

Property Value

double?

Remarks

For Beginners: A value between 0 and 1 indicating how confident the model is in its prediction. Higher = more confident. May be null if the model doesn't provide confidence scores.

IndividualPredictions

Gets all the individual predictions from each augmented version.

public IReadOnlyList<TOutput> IndividualPredictions { get; }

Property Value

IReadOnlyList<TOutput>

Remarks

For Beginners: This shows what the model predicted for each variation. Useful for: - Debugging: See if one augmentation is causing problems - Analysis: Understand how much predictions vary - Visualization: Show uncertainty in results

StandardDeviation

Gets the standard deviation of predictions, measuring uncertainty.

public double? StandardDeviation { get; }

Property Value

double?

Remarks

For Beginners: This tells you how much the predictions varied: - Low standard deviation: Predictions were consistent (good!) - High standard deviation: Predictions varied a lot (less reliable)

Use this to decide how much to trust the result.