TrialResult

Models & Types · AiDotNet.AutoML

Represents the result of a single trial during AutoML search.

For Beginners

AutoML runs many small experiments called “trials”. Each trial:

  • Tries one model family with one set of settings.
  • Trains the model and scores it.
  • Records whether it succeeded and how well it did.

This class stores the outcome of one of those experiments.

How It Works

This type is used by internal AutoML implementations to track trial execution outcomes. When accessed via public APIs (for example, through GetTrialHistory), sensitive fields like raw hyperparameter values must be redacted to align with the AiDotNet facade/IP goals.

Constructors

Constructor Summary
TrialResult()

Properties

Property Summary
DateTime Timestamp { get; set; } Gets or sets the UTC timestamp when the trial was completed.
Dictionary<string, object> Metadata { get; set; } Gets or sets additional metadata about the trial.
Dictionary<string, object> Parameters { get; set; } Gets or sets the hyperparameters used in this trial.
TimeSpan Duration { get; set; } Gets or sets the duration of the trial.
Type CandidateModelType { get; set; } Gets or sets the candidate model family used for the trial, when known.
bool Success { get; set; } Gets or sets a value indicating whether the trial completed successfully.
double Score { get; set; } Gets or sets the score achieved by this trial.
int TrialId { get; set; } Gets or sets the unique identifier for the trial.
string ErrorMessage { get; set; } Gets or sets an error message if the trial failed.

Methods

Method Summary
TrialResult Clone() Creates a deep copy of this trial result including sensitive parameters.
TrialResult CloneRedacted() Creates a deep copy of this trial result with sensitive fields redacted.