private void ExecuteStartTraining(bool canEvaluate)
{
((MainWindowVM)Parent).Save();
var config = GetTestConfiguration();
((MainWindowVM)Parent).ExpressionsVisibility = Visibility.Hidden;
try
{
if (_defaultChooser == null)
{
_defaultChooser = new DefaultExpressionChooser(DateTimeProvider, DataAccess);
_flashCardsAlgorithmChooser = new FlashCardsExpressionsChooser(DateTimeProvider, DataAccess);
_defaultLearningAlgorithm = new DefaultLearningAlgorithm(DateTimeProvider);
_flashCardsLearningAlgorithm = new FlashCardsLearningAlgorithm();
}
WindowService.ShowDialog(
new TestVM(
canEvaluate,
config,
FlashCardsAlgorithm ? _flashCardsAlgorithmChooser : _defaultChooser,
FlashCardsAlgorithm ? _flashCardsLearningAlgorithm : _defaultLearningAlgorithm));
}
catch (Exception ex)
{
WindowService.ShowError(ex);
}
((MainWindowVM)Parent).ExpressionsVisibility = Visibility.Visible;
((MainWindowVM)Parent).RebuildList();
}
public TestVM(bool canEvaluate, TestConfiguration config, IExpressionsChooser chooser, ILearningAlgorithm algorithm)
{
config.NotNull("config");
chooser.NotNull("chooser");
algorithm.NotNull("algorithm");
CanEvaluate = canEvaluate;
Config = config;
Algorithm = algorithm;
Expressions = chooser.SelectExpressions(Config);
ExpressionsView = CollectionViewSource.GetDefaultView(Expressions);
}
/// <summary>
/// Initializes a new instance of the <see cref="MultiLayerNetwork{T}" /> class.
/// </summary>
/// <param name="learningAlgorithm">The learning algorithm.</param>
/// <param name="minWeight">The minimum weight.</param>
/// <param name="maxWeight">The maximum weight.</param>
/// <param name="weightProbabilities">The weight probability distribution.</param>
/// <param name="maxInputActivations">The maximum number of activated input units.</param>
/// <param name="activationThreshold">The total 'signal' required for a unit to become active.</param>
/// <param name="classes">The number of distinct classes.</param>
/// <param name="unitsPerClass">The number of output units per class.</param>
/// <param name="units">The number of units in each layer, excluding the output layer.</param>
public MultiLayerNetwork(
ILearningAlgorithm learningAlgorithm,
int minWeight,
int maxWeight,
double[] weightProbabilities,
int maxInputActivations,
int activationThreshold,
int classes,
int unitsPerClass,
params int[] units)
: base(
learningAlgorithm,
minWeight,
maxWeight,
weightProbabilities,
1.0, // Synapse success probability
maxInputActivations,
activationThreshold,
classes,
unitsPerClass,
units)
{
}
/// <summary>
/// Initializes the agent
/// </summary>
/// <param name="environment">The environment that the agent is added to</param>
public void Initialize(TEnvironment environment)
{
this.environment = environment;
this.learner = this.createLearnerFn(environment);
}
/// <summary>
/// Initializes the agent
/// </summary>
/// <param name="environment">The environment that the agent is added to</param>
public virtual void Initialize(TEnvironment environment)
{
this.environment = environment;
this.learner = this.CreateLearner();
}
