public int DetermineTreeType(OutputEquilateral eqField, IMLData output)
{
int result;
if (eqField != null)
{
result = eqField.Equilateral.Decode(output.Data);
}
else
{
double maxOutput = Double.NegativeInfinity;
result = -1;
for (int i = 0; i < output.Count; i++)
{
if (output[i] > maxOutput)
{
maxOutput = output[i];
result = i;
}
}
}
return result;
}
/// <summary>
/// Calculate the best matching unit (BMU). This is the output neuron that
/// has the lowest Euclidean distance to the input vector.
/// </summary>
///
/// <param name="input">The input vector.</param>
/// <returns>The output neuron number that is the BMU.</returns>
public int CalculateBMU(IMLData input)
{
int result = 0;
// Track the lowest distance so far.
double lowestDistance = Double.MaxValue;
for (int i = 0; i < _som.OutputCount; i++)
{
double distance = CalculateEuclideanDistance(
_som.Weights, input, i);
// Track the lowest distance, this is the BMU.
if (distance < lowestDistance)
{
lowestDistance = distance;
result = i;
}
}
// Track the worst distance, this is the error for the entire network.
if (lowestDistance > _worstDistance)
{
_worstDistance = lowestDistance;
}
return result;
}
/// <inheritdoc/>
public void CalculateError(IMLData ideal, double[] actual, double[] error)
{
for (int i = 0; i < actual.Length; i++)
{
error[i] = ideal[i] - actual[i];
}
}
/// <summary>
/// Classify the input into one of the output clusters.
/// </summary>
/// <param name="input">The input.</param>
/// <returns>The cluster it was clasified into.</returns>
public int Classify(IMLData input)
{
if (input.Count > InputCount)
{
throw new NeuralNetworkError(
"Can't classify SOM with input size of " + InputCount
+ " with input data of count " + input.Count);
}
double[][] m = _weights.Data;
double minDist = Double.PositiveInfinity;
int result = -1;
for (int i = 0; i < OutputCount; i++)
{
double dist = EngineArray.EuclideanDistance(input, m[i]);
if (dist < minDist)
{
minDist = dist;
result = i;
}
}
return result;
}
public override sealed IMLData Compute(IMLData input)
{
int num;
BiPolarMLData data = new BiPolarMLData(input.Count);
if (0 == 0)
{
if (((uint) num) <= uint.MaxValue)
{
if (3 == 0)
{
return data;
}
goto Label_0053;
}
}
else
{
goto Label_0053;
}
Label_003B:
EngineArray.ArrayCopy(base.CurrentState.Data, data.Data);
return data;
Label_0053:
EngineArray.ArrayCopy(input.Data, base.CurrentState.Data);
this.Run();
for (num = 0; num < base.CurrentState.Count; num++)
{
data.SetBoolean(num, BiPolarUtil.Double2bipolar(base.CurrentState[num]));
}
goto Label_003B;
}
private void x3342cd5bc15ae07b(int x7079b5ea66d0bae1, IMLData xcdaeea7afaf570ff)
{
for (int i = 0; i < this._x87a7fc6a72741c2e.InputCount; i++)
{
this._x87a7fc6a72741c2e.Weights[i, x7079b5ea66d0bae1] = xcdaeea7afaf570ff[i];
}
}
/// <summary>
/// Copy the specified input pattern to the weight matrix. This causes an
/// output neuron to learn this pattern "exactly". This is useful when a
/// winner is to be forced.
/// </summary>
///
/// <param name="outputNeuron">The output neuron to set.</param>
/// <param name="input">The input pattern to copy.</param>
private void CopyInputPattern(int outputNeuron, IMLData input)
{
for (int inputNeuron = 0; inputNeuron < _network.InputCount; inputNeuron++)
{
_network.Weights[inputNeuron, outputNeuron] = input[inputNeuron];
}
}
public IMLData Compute(IMLData input)
{
int num;
Matrix col;
Matrix matrix2;
IMLData data = new BasicMLData(this.OutputCount);
if (0 == 0)
{
goto Label_003F;
}
Label_000F:
matrix2 = Matrix.CreateRowMatrix(input.Data);
if (3 == 0)
{
goto Label_003F;
}
data[num] = MatrixMath.DotProduct(matrix2, col);
num++;
Label_0034:
if (num < this.OutputCount)
{
col = this._weights.GetCol(num);
goto Label_000F;
}
return data;
Label_003F:
num = 0;
goto Label_0034;
}
/// <summary>
/// Train the neural network for the specified pattern. The neural network
/// can be trained for more than one pattern. To do this simply call the
/// train method more than once.
/// </summary>
///
/// <param name="pattern">The pattern to train for.</param>
public void AddPattern(IMLData pattern)
{
if (pattern.Count != NeuronCount)
{
throw new NeuralNetworkError("Network with " + NeuronCount
+ " neurons, cannot learn a pattern of size "
+ pattern.Count);
}
// Create a row matrix from the input, convert boolean to bipolar
Matrix m2 = Matrix.CreateRowMatrix(pattern.Data);
// Transpose the matrix and multiply by the original input matrix
Matrix m1 = MatrixMath.Transpose(m2);
Matrix m3 = MatrixMath.Multiply(m1, m2);
// matrix 3 should be square by now, so create an identity
// matrix of the same size.
Matrix identity = MatrixMath.Identity(m3.Rows);
// subtract the identity matrix
Matrix m4 = MatrixMath.Subtract(m3, identity);
// now add the calculated matrix, for this pattern, to the
// existing weight matrix.
ConvertHopfieldMatrix(m4);
}
public BasicMLDataPair(IMLData input, IMLData ideal, IMLData calced, IMLData error)
{
this._significance = 1.0;
this._input = input;
this._ideal = ideal;
this._calced = calced;
this._error = error;
}
public override void Add(IMLData data)
{
if (!(data is ImageMLData))
{
throw new NeuralNetworkError("This data set only supports ImageNeuralData or Image objects.");
}
base.Add(data);
}
public BasicMLDataPair(IMLData input, IMLData ideal)
{
this._significance = 1.0;
this._input = input;
this._ideal = ideal;
this._calced = null;
this._error = null;
}
public void UpdateError(double[] actual, IMLData ideal)
{
for (int i = 0; i < actual.Length; i++)
{
double delta = ideal[i] - actual[i];
_globalError += delta * delta;
}
_setSize += ideal.Count;
}
/// <summary>
/// Construct a loaded row from an IMLData.
/// </summary>
/// <param name="format">The format to store the numbers in.</param>
/// <param name="data">The data to use.</param>
/// <param name="extra">The extra positions to allocate.</param>
public LoadedRow(CSVFormat format, IMLData data, int extra)
{
int count = data.Count;
_data = new String[count + extra];
for (int i = 0; i < count; i++)
{
_data[i] = format.Format(data[i], 5);
}
}
/// <summary>
/// Add the specified input and ideal object to the collection.
/// </summary>
/// <param name="inputData">The image to train with.</param>
/// <param name="idealData">The expected otuput form this image.</param>
public override void Add(IMLData inputData, IMLData idealData)
{
if (!(inputData is ImageMLData))
{
throw new NeuralNetworkError(MUST_USE_IMAGE);
}
base.Add(inputData, idealData);
}
/// <summary>
/// Subtract two vectors.
/// </summary>
/// <param name="a">First vector.</param>
/// <param name="b">Second vector.</param>
/// <returns>Result vector.</returns>
public static double[] Subtract(IMLData a, double[] b)
{
double[] result = new double[a.Count];
for (int i = 0; i < a.Count; i++)
{
result[i] = a[i] - b[i];
}
return(result);
}
public void CalculateError(IActivationFunction af, double[] b, double[] a,
IMLData ideal, double[] actual, double[] error, double derivShift,
double significance)
{
for (int i = 0; i < actual.Length; i++)
{
error[i] = (ideal[i] - actual[i]) * significance;
}
}
/// <summary>
/// Add the specified data, must be an ImageNeuralData class.
/// </summary>
/// <param name="data">The data The object to add.</param>
public override void Add(IMLData data)
{
if (!(data is ImageMLData))
{
throw new NeuralNetworkError(MUST_USE_IMAGE);
}
base.Add(data);
}
/// <summary>
/// Get the specified weight.
/// </summary>
///
/// <param name="matrix">The matrix to use.</param>
/// <param name="input">The input, to obtain the size from.</param>
/// <param name="x"></param>
/// <param name="y"></param>
/// <returns>The value from the matrix.</returns>
private static double GetWeight(Matrix matrix, IMLData input,
int x, int y)
{
if (matrix.Rows != input.Count)
{
return(matrix[x, y]);
}
return(matrix[y, x]);
}
public int CalculateScore(IEnumerable <int> customerOrder, bool showResults)
{
var input = new BasicMLData(1);
input[0] = 1;
IMLData output = _network.Compute(input);
var logOutput = new List <LogisticSimulatorOutput>();
logOutput.Add(new LogisticSimulatorOutput()
{
Name = "Retailer_Min", Value = Convert.ToInt32(min.DeNormalize(output[0]))
});
logOutput.Add(new LogisticSimulatorOutput()
{
Name = "Retailer_Max", Value = Convert.ToInt32(max.DeNormalize(output[1]))
});
logOutput.Add(new LogisticSimulatorOutput()
{
Name = "WholeSaler_Min", Value = Convert.ToInt32(min.DeNormalize(output[2]))
});
logOutput.Add(new LogisticSimulatorOutput()
{
Name = "WholeSaler_Max", Value = Convert.ToInt32(max.DeNormalize(output[3]))
});
logOutput.Add(new LogisticSimulatorOutput()
{
Name = "Distributor_Min", Value = Convert.ToInt32(min.DeNormalize(output[4]))
});
logOutput.Add(new LogisticSimulatorOutput()
{
Name = "Distributor_Max", Value = Convert.ToInt32(max.DeNormalize(output[5]))
});
logOutput.Add(new LogisticSimulatorOutput()
{
Name = "Factory_Min", Value = Convert.ToInt32(min.DeNormalize(output[6]))
});
logOutput.Add(new LogisticSimulatorOutput()
{
Name = "Factory_Max", Value = Convert.ToInt32(max.DeNormalize(output[7]))
});
logOutput.Add(new LogisticSimulatorOutput()
{
Name = "Factory_Units_Per_Day", Value = Convert.ToInt32(units.DeNormalize(output[8]))
});
sim.Start(logOutput, customOrders: customerOrder);
if (showResults)
{
ShowResults(logOutput);
}
int score = Convert.ToInt32(sim.SumAllCosts());
return(score);
}
private char Denormalize(IMLData value)
{
var query = new double[value.Count];
value.CopyTo(query, 0, value.Count);
var index = query.ToList().IndexOf(1D);
return(_alphas[index]);
}
/// <summary>
/// Note: for Boltzmann networks, you will usually want to call the "run"
/// method to compute the output.
/// This method can be used to copy the input data to the current state. A
/// single iteration is then run, and the new current state is returned.
/// </summary>
///
/// <param name="input">The input pattern.</param>
/// <returns>The new current state.</returns>
public override sealed IMLData Compute(IMLData input)
{
var result = new BiPolarMLData(input.Count);
input.CopyTo(CurrentState.Data, 0, input.Count);
Run();
EngineArray.ArrayCopy(CurrentState.Data, result.Data);
return(result);
}
public virtual void Compute(IMLData input, double[] output)
{
int sourceIndex = _layerOutput.Length
- _layerCounts[_layerCounts.Length - 1];
input.CopyTo(_layerOutput, sourceIndex, _inputCount);
InnerCompute(output);
}
static void Main(string[] args)
{
// used for prediction of time series
// sin(x) in theory
int DEGREES = 360;
int WINDOW_SIZE = 16;
double[][] Input = new double[DEGREES][];
double[][] Ideal = new double[DEGREES][];
// Create array of sin signals
for (int i = 0; i < DEGREES; i++)
{
Input[i] = new double[WINDOW_SIZE];
Ideal[i] = new double[] { Math.Sin(DegreeToRad(i + WINDOW_SIZE)) };
for (int j = 0; j < WINDOW_SIZE; j++)
{
Input[i][j] = Math.Sin(DegreeToRad(i + j));
}
}
// construct training set
IMLDataSet trainingSet = new BasicMLDataSet(Input, Ideal);
// construct an Elman type network
// simple recurrent network
ElmanPattern pattern = new ElmanPattern
{
InputNeurons = WINDOW_SIZE,
ActivationFunction = new ActivationSigmoid(),
OutputNeurons = 1
};
pattern.AddHiddenLayer(WINDOW_SIZE);
IMLMethod method = pattern.Generate();
BasicNetwork network = (BasicNetwork)method;
// Train network
IMLTrain train = new Backpropagation(network, trainingSet);
var stop = new StopTrainingStrategy();
train.AddStrategy(new Greedy());
train.AddStrategy(stop);
int epoch = 0;
while (!stop.ShouldStop())
{
train.Iteration();
Console.WriteLine($"Training Epoch #{epoch} Error:{train.Error}");
epoch++;
}
// Test network
foreach (IMLDataPair pair in trainingSet)
{
IMLData output = network.Compute(pair.Input);
Console.WriteLine($"actual={output[0]}, ideal={pair.Ideal[0]}");
}
}
/// <summary>
/// Add only input data, for an unsupervised dataset.
/// </summary>
/// <param name="data1">The data to be added.</param>
public void Add(IMLData data1)
{
if (!_loading)
{
throw new IMLDataError(ErrorAdd);
}
_egb.Write(data1);
_egb.Write(1.0);
}
public void CalculateError(IActivationFunction af, double[] b, double[] a,
IMLData ideal, double[] actual, double[] error, double derivShift,
double significance)
{
for (int i = 0; i < actual.Length; i++)
{
double deriv = af.DerivativeFunction(b[i], a[i]);
error[i] = (Math.Atan(ideal[i] - actual[i]) * significance) * deriv;
}
}
private double[] IMDataToDoubleArray(IMLData d)
{
double[] array = new double[d.Count];
for (int i = 0; i < d.Count; i++)
{
array[i] = d[i];
}
return(array);
}
/// <summary>
/// Construct a loaded row from an IMLData.
/// </summary>
/// <param name="format">The format to store the numbers in.</param>
/// <param name="data">The data to use.</param>
/// <param name="extra">The extra positions to allocate.</param>
public LoadedRow(CSVFormat format, IMLData data, int extra)
{
int count = data.Count;
_data = new String[count + extra];
for (int i = 0; i < count; i++)
{
_data[i] = format.Format(data[i], 5);
}
}
/// <summary>
/// Add only input data, for an unsupervised dataset.
/// </summary>
/// <param name="data1">The data to be added.</param>
public void Add(IMLData data1)
{
if (!loading)
{
throw new IMLDataError(ERROR_ADD);
}
egb.Write(data1.Data);
egb.Write(1.0);
}
/// <summary>
/// Add both the input and ideal data.
/// </summary>
/// <param name="inputData">The input data.</param>
/// <param name="idealData">The ideal data.</param>
public void Add(IMLData inputData, IMLData idealData)
{
if (!loading)
{
throw new IMLDataError(ERROR_ADD);
}
egb.Write(inputData.Data);
egb.Write(idealData.Data);
egb.Write(1.0);
}
/// <summary>
/// Get the Euclidean distance between the specified data and the set number.
/// </summary>
/// <param name="data">The data to check.</param>
/// <param name="set">The set to check.</param>
/// <returns>The distance.</returns>
public double GetDistance(IMLData data, int set)
{
double result = 0;
for (int i = 0; i < data.Count; i++)
{
var val = data[i] - _matrix[set][i];
result += val * val;
}
return(Math.Sqrt(result));
}
/// <summary>
/// Add both the input and ideal data.
/// </summary>
/// <param name="inputData">The input data.</param>
/// <param name="idealData">The ideal data.</param>
public void Add(IMLData inputData, IMLData idealData)
{
if (!_loading)
{
throw new IMLDataError(ErrorAdd);
}
_egb.Write(inputData);
_egb.Write(idealData);
_egb.Write(1.0);
}
public double PredictNext(double[] inputData)
{
double[] newNormalizedData = normalizeArray.Process(inputData);
IMLData input = new BasicMLData(WindowSize);
input.Data = newNormalizedData;
IMLData output1 = network.Compute(input);
double outputDenormalized = normalizeArray.Stats.DeNormalize(output1.Data[0]);
return(outputDenormalized);
}
/// <summary>
/// Called to update for each number that should be checked.
/// </summary>
public void UpdateError(IMLData actual, double[] ideal, double significance)
{
for (int i = 0; i < actual.Count; i++)
{
double delta = ideal[i] - actual[i];
_globalError += delta * delta;
}
_setSize += ideal.Length;
}
public double DetermineAngle(IMLData angle)
{
double result = (Math.Atan2(angle[0], angle[1]) / Math.PI) * 180;
if (result < 0)
{
result += 360;
}
return(result);
}
/// <summary>
/// Calculate the Euclidean distance between two vectors.
/// </summary>
/// <param name="p1">The first vector.</param>
/// <param name="p2">The second vector.</param>
/// <returns>The distance.</returns>
public static double EuclideanDistance(IMLData p1, double[] p2)
{
double sum = 0;
for (int i = 0; i < p1.Count; i++)
{
double d = p1[i] - p2[i];
sum += d * d;
}
return(Math.Sqrt(sum));
}
private string GetOutput(IMLData lst)
{
string ret = " [Ouput: ";
for (int i = 0; i < lst.Count; i++)
{
ret += lst[i] + ",";
}
ret = ret.TrimEnd(',');
ret += "]";
return(ret);
}
protected override void Update(GameTime gameTime)
{
if (GamePad.GetState(PlayerIndex.One).Buttons.Back == ButtonState.Pressed || Keyboard.GetState().IsKeyDown(Keys.Escape))
{
Exit();
}
var input = debugGame.inputs();
IMLData output = net.Compute(input);
debugGame.update(output[0], output[1]);
}
// Predict data using a network.
public static bool Predict(ref NetworkContainer container, ref EncogPredictSettings settings)
{
IMLData output = container.network.Compute(new BasicMLData(settings.data));
if (container.verbose)
{
Console.WriteLine("Raw output: " + output[0]);
}
// Return true or false according to threshold.
return(output[0] >= settings.threshold);
}
/// <inheritdoc/>
public double Distance(IMLDataPair d)
{
IMLData diff = _value.Minus(d.Input);
double sum = 0.0;
for (int i = 0; i < diff.Count; i++)
{
sum += diff[i] * diff[i];
}
return(Math.Sqrt(sum));
}
/// <inheritdoc/>
public int Classify(IMLData input)
{
if (_model == null)
{
throw new EncogError(
"Can't use the SVM yet, it has not been trained, "
+ "and no model exists.");
}
svm_node[] formattedInput = MakeSparse(input);
return((int)svm.svm_predict(_model, formattedInput));
}
/// <summary>
/// Process the file.
/// </summary>
/// <param name="outputFile">The output file.</param>
/// <param name="method">The method to use.</param>
public void Process(FileInfo outputFile, IMLRegression method)
{
var csv = new ReadCSV(InputFilename.ToString(),
ExpectInputHeaders, Format);
if (method.InputCount != _inputCount)
{
throw new AnalystError("This machine learning method has "
+ method.InputCount
+ " inputs, however, the data has " + _inputCount
+ " inputs.");
}
var input = new BasicMLData(method.InputCount);
StreamWriter tw = AnalystPrepareOutputFile(outputFile);
ResetStatus();
while (csv.Next())
{
UpdateStatus(false);
var row = new LoadedRow(csv, _idealCount);
int dataIndex = 0;
// load the input data
for (int i = 0; i < _inputCount; i++)
{
String str = row.Data[i];
double d = Format.Parse(str);
input[i] = d;
dataIndex++;
}
// do we need to skip the ideal values?
dataIndex += _idealCount;
// compute the result
IMLData output = method.Compute(input);
// display the computed result
for (int i = 0; i < _outputCount; i++)
{
double d = output[i];
row.Data[dataIndex++] = Format.Format(d, Precision);
}
WriteRow(tw, row);
}
ReportDone(false);
tw.Close();
csv.Close();
}
/// <summary>
/// Add a pattern to the neural network.
/// </summary>
///
/// <param name="inputPattern">The input pattern.</param>
/// <param name="outputPattern">The output pattern(for this input).</param>
public void AddPattern(IMLData inputPattern,
IMLData outputPattern)
{
for (int i = 0; i < _f1Count; i++)
{
for (int j = 0; j < _f2Count; j++)
{
var weight = (int)(inputPattern[i] * outputPattern[j]);
_weightsF1ToF2.Add(i, j, weight);
_weightsF2ToF1.Add(j, i, weight);
}
}
}
/// <inheritdoc />
public String DenormalizeColumn(ColumnDefinition colDef, IMLData data,
int dataColumn)
{
double value = data[dataColumn];
double result = ((colDef.Low - colDef.High)*value
- _normalizedHigh*colDef.Low + colDef.High
*_normalizedLow)
/(_normalizedLow - _normalizedHigh);
// typically caused by a number that should not have been normalized
// (i.e. normalization or actual range is infinitely small.
if (Double.IsNaN(result))
{
return "" + (((_normalizedHigh - _normalizedLow)/2) + _normalizedLow);
}
return "" + result;
}
/// <inheritdoc />
public String DenormalizeColumn(ColumnDefinition colDef, IMLData data,
int dataColumn)
{
double bestValue = Double.NegativeInfinity;
int bestIndex = 0;
for (int i = 0; i < data.Count; i++)
{
double d = data[dataColumn + i];
if (d > bestValue)
{
bestValue = d;
bestIndex = i;
}
}
return colDef.Classes[bestIndex];
}
/// <summary>
/// Construct a new BasicMLData object from an existing one. This makes a
/// copy of an array. If MLData is not complex, then only reals will be
/// created.
/// </summary>
/// <param name="d">The object to be copied.</param>
public BasicMLComplexData(IMLData d)
{
if (d is IMLComplexData)
{
var c = (IMLComplexData) d;
for (int i = 0; i < d.Count; i++)
{
_data[i] = new ComplexNumber(c.GetComplexData(i));
}
}
else
{
for (int i = 0; i < d.Count; i++)
{
_data[i] = new ComplexNumber(d[i], 0);
}
}
}
/// <inheritdoc />
public String DenormalizeColumn(ColumnDefinition colDef, IMLData data,
int dataColumn)
{
double high = colDef.Classes.Count;
double low = 0;
double value = data[dataColumn];
double result = ((low - high)*value - _normalizedHigh*low + high
*_normalizedLow)
/(_normalizedLow - _normalizedHigh);
// typically caused by a number that should not have been normalized
// (i.e. normalization or actual range is infinitely small.
if (Double.IsNaN(result))
{
return colDef.Classes[0];
}
return colDef.Classes[(int) result];
}
public void AddPattern(IMLData pattern)
{
object[] objArray;
if (pattern.Count != base.NeuronCount)
{
objArray = new object[4];
if (0 == 0)
{
objArray[0] = "Network with ";
do
{
if (0 != 0)
{
return;
}
}
while (0 != 0);
objArray[1] = base.NeuronCount;
objArray[2] = " neurons, cannot learn a pattern of size ";
objArray[3] = pattern.Count;
}
}
else
{
Matrix b = Matrix.CreateRowMatrix(pattern.Data);
Matrix a = MatrixMath.Multiply(MatrixMath.Transpose(b), b);
Matrix matrix4 = MatrixMath.Identity(a.Rows);
Matrix delta = MatrixMath.Subtract(a, matrix4);
this.ConvertHopfieldMatrix(delta);
if (0 == 0)
{
return;
}
}
throw new NeuralNetworkError(string.Concat(objArray));
}
private double MaxInArray(IMLData result)
{
var array = new double[Network.OutputSize];
result.CopyTo(array, 0, Network.OutputSize - 1);
var maxValue = array.Max();
var maxIndex = array.ToList().IndexOf(maxValue);
return maxIndex + 1;
}
/// <summary>
/// Compute the output from this synapse.
/// </summary>
/// <param name="input">The input to this synapse.</param>
/// <returns>The output from this synapse.</returns>
public IMLData Compute(IMLData input)
{
var result = new double[_outputCount];
// clear from previous
EngineArray.Fill(_preActivation, 0.0);
EngineArray.Fill(_postActivation, 0.0);
_postActivation[0] = 1.0;
// copy input
for (int i = 0; i < _inputCount; i++)
{
_postActivation[i + 1] = input[i];
}
// iterate through the network activationCycles times
for (int i = 0; i < ActivationCycles; ++i)
{
InternalCompute();
}
// copy output
for (int i = 0; i < _outputCount; i++)
{
result[i] = _postActivation[_outputIndex + i];
}
return new BasicMLData(result);
}
/// <summary>
/// Create a matrix that is a single row.
/// </summary>
/// <param name="input">A 1D array to make the matrix from.</param>
/// <returns>A matrix that contans a single row.</returns>
public static Matrix CreateRowMatrix(IMLData input)
{
var d = new double[1][];
d[0] = new double[input.Count];
for(int i = 0; i < input.Count; i++)
{
d[0][i] = input[i];
}
return new Matrix(d);
}
/// <summary>
/// Not supported.
/// </summary>
/// <param name="inputData">Not used.</param>
/// <param name="idealData">Not used.</param>
public void Add(IMLData inputData, IMLData idealData)
{
throw new TrainingError(AddNotSupported);
}
/// <summary>
/// Not supported.
/// </summary>
/// <param name="data1">Not used.</param>
public void Add(IMLData data1)
{
throw new TrainingError(AddNotSupported);
}
/// <summary>
/// Determines the angle on IMLData.
/// </summary>
/// <param name="angle">The angle.</param>
/// <returns></returns>
public static double DetermineAngle(IMLData angle)
{
double result = (Math.Atan2(angle[0], angle[1])/Math.PI)*180;
if (result < 0)
result += 360;
return result;
}
/// <summary>
/// Compute the output from this network.
/// </summary>
///
/// <param name="input">The input to the network.</param>
/// <returns>The output from the network.</returns>
public override sealed IMLData Compute(IMLData input)
{
var xout = new double[OutputCount];
double psum = 0.0d;
int r = -1;
foreach (IMLDataPair pair in _samples)
{
r++;
if (r == Exclude)
{
continue;
}
double dist = 0.0d;
for (int i = 0; i < InputCount; i++)
{
double diff = input[i] - pair.Input[i];
diff /= _sigma[i];
dist += diff*diff;
}
if (Kernel == PNNKernelType.Gaussian)
{
dist = Math.Exp(-dist);
}
else if (Kernel == PNNKernelType.Reciprocal)
{
dist = 1.0d/(1.0d + dist);
}
if (dist < 1.0e-40d)
{
dist = 1.0e-40d;
}
if (OutputMode == PNNOutputMode.Classification)
{
var pop = (int) pair.Ideal[0];
xout[pop] += dist;
}
else if (OutputMode == PNNOutputMode.Unsupervised)
{
for (int i = 0; i < InputCount; i++)
{
xout[i] += dist*pair.Input[i];
}
psum += dist;
}
else if (OutputMode == PNNOutputMode.Regression)
{
for (int i = 0; i < OutputCount; i++)
{
xout[i] += dist*pair.Ideal[i];
}
psum += dist;
}
}
if (OutputMode == PNNOutputMode.Classification)
{
psum = 0.0d;
for (int i = 0; i < OutputCount; i++)
{
if (_priors[i] >= 0.0d)
{
xout[i] *= _priors[i]/_countPer[i];
}
psum += xout[i];
}
if (psum < 1.0e-40d)
{
psum = 1.0e-40d;
}
for (int i = 0; i < OutputCount; i++)
{
xout[i] /= psum;
}
}
else if (OutputMode == PNNOutputMode.Unsupervised)
{
for (int i = 0; i < InputCount; i++)
{
xout[i] /= psum;
}
}
else if (OutputMode == PNNOutputMode.Regression)
{
for (int i = 0; i < OutputCount; i++)
{
xout[i] /= psum;
}
}
return new BasicMLData(xout);
}
/// <inheritdoc/>
public int Classify(IMLData input)
{
IMLData output = Compute(input);
return EngineArray.MaxIndex(output.Data);
}
/// <summary>
/// Called to update for each number that should be checked.
/// </summary>
public void UpdateError(IMLData actual, double[] ideal, double significance)
{
for(int i = 0; i < actual.Count; i++)
{
double delta = ideal[i] - actual[i];
_globalError += delta * delta;
}
_setSize += ideal.Length;
}
/// <summary>
/// Compute the output from the input MLData. The individual values of the
/// input will be mapped to the variables defined in the context. The order
/// is the same between the input and the defined variables. The input will
/// be mapped to the appropriate types. Enums will use their ordinal number.
/// The result will be a single number MLData.
/// </summary>
/// <param name="input">The input to the program.</param>
/// <returns>A single numer MLData.</returns>
public IMLData Compute(IMLData input)
{
if (input.Count != InputCount)
{
throw new EACompileError("Invalid input count.");
}
for (int i = 0; i < input.Count; i++)
{
_variables.SetVariable(i, input[i]);
}
ExpressionValue v = RootNode.Evaluate();
VariableMapping resultMapping = ResultType;
var result = new BasicMLData(1);
bool success = false;
switch (resultMapping.VariableType)
{
case EPLValueType.FloatingType:
if (v.IsNumeric)
{
result.Data[0] = v.ToFloatValue();
success = true;
}
break;
case EPLValueType.StringType:
result.Data[0] = v.ToFloatValue();
success = true;
break;
case EPLValueType.BooleanType:
if (v.IsBoolean)
{
result.Data[0] = v.ToBooleanValue() ? 1.0 : 0.0;
success = true;
}
break;
case EPLValueType.IntType:
if (v.IsNumeric)
{
result[0] = v.ToIntValue();
success = true;
}
break;
case EPLValueType.EnumType:
if (v.IsEnum)
{
result.Data[0] = v.ToIntValue();
success = true;
}
break;
}
if (!success)
{
throw new EARuntimeError("EncogProgram produced "
+ v.ExprType.ToString() + " but "
+ resultMapping.VariableType.ToString()
+ " was expected.");
}
return result;
}
/// <summary>
/// Write the data from an IMLData
/// </summary>
/// <param name="v">The array to write.</param>
public void Write(IMLData v)
{
try
{
for(int i = 0; i < v.Count; i++)
{
_binaryWriter.Write(v[i]);
}
}
catch(IOException ex)
{
throw new BufferedDataError(ex);
}
}
/// <summary>
/// Compute the derivative for target data.
/// </summary>
///
/// <param name="input">The input.</param>
/// <param name="target">The target data.</param>
/// <returns>The output.</returns>
public IMLData ComputeDeriv(IMLData input, IMLData target)
{
int pop, ivar;
int ibest = 0;
int outvar;
double dist, truedist;
double vtot, wtot;
double temp, der1, der2, psum;
int vptr, wptr, vsptr = 0, wsptr = 0;
var xout = new double[_network.OutputCount];
for (pop = 0; pop < _network.OutputCount; pop++)
{
xout[pop] = 0.0d;
for (ivar = 0; ivar < _network.InputCount; ivar++)
{
_v[pop*_network.InputCount + ivar] = 0.0d;
_w[pop*_network.InputCount + ivar] = 0.0d;
}
}
psum = 0.0d;
if (_network.OutputMode != PNNOutputMode.Classification)
{
vsptr = _network.OutputCount
*_network.InputCount;
wsptr = _network.OutputCount
*_network.InputCount;
for (ivar = 0; ivar < _network.InputCount; ivar++)
{
_v[vsptr + ivar] = 0.0d;
_w[wsptr + ivar] = 0.0d;
}
}
IMLDataPair pair;
for (int r = 0; r < _network.Samples.Count; r++)
{
pair = _network.Samples[r];
if (r == _network.Exclude)
{
continue;
}
dist = 0.0d;
for (ivar = 0; ivar < _network.InputCount; ivar++)
{
double diff = input[ivar] - pair.Input[ivar];
diff /= _network.Sigma[ivar];
_dsqr[ivar] = diff*diff;
dist += _dsqr[ivar];
}
if (_network.Kernel == PNNKernelType.Gaussian)
{
dist = Math.Exp(-dist);
}
else if (_network.Kernel == PNNKernelType.Reciprocal)
{
dist = 1.0d/(1.0d + dist);
}
truedist = dist;
if (dist < 1.0e-40d)
{
dist = 1.0e-40d;
}
if (_network.OutputMode == PNNOutputMode.Classification)
{
pop = (int) pair.Ideal[0];
xout[pop] += dist;
vptr = pop*_network.InputCount;
wptr = pop*_network.InputCount;
for (ivar = 0; ivar < _network.InputCount; ivar++)
{
temp = truedist*_dsqr[ivar];
_v[vptr + ivar] += temp;
_w[wptr + ivar] += temp*(2.0d*_dsqr[ivar] - 3.0d);
}
}
else if (_network.OutputMode == PNNOutputMode.Unsupervised)
{
for (ivar = 0; ivar < _network.InputCount; ivar++)
{
xout[ivar] += dist*pair.Input[ivar];
temp = truedist*_dsqr[ivar];
_v[vsptr + ivar] += temp;
_w[wsptr + ivar] += temp
*(2.0d*_dsqr[ivar] - 3.0d);
}
vptr = 0;
wptr = 0;
for (outvar = 0; outvar < _network.OutputCount; outvar++)
{
for (ivar = 0; ivar < _network.InputCount; ivar++)
{
temp = truedist*_dsqr[ivar]
*pair.Input[ivar];
_v[vptr++] += temp;
_w[wptr++] += temp*(2.0d*_dsqr[ivar] - 3.0d);
}
}
psum += dist;
}
else if (_network.OutputMode == PNNOutputMode.Regression)
{
for (ivar = 0; ivar < _network.OutputCount; ivar++)
{
xout[ivar] += dist*pair.Ideal[ivar];
}
vptr = 0;
wptr = 0;
for (outvar = 0; outvar < _network.OutputCount; outvar++)
{
for (ivar = 0; ivar < _network.InputCount; ivar++)
{
temp = truedist*_dsqr[ivar]
*pair.Ideal[outvar];
_v[vptr++] += temp;
_w[wptr++] += temp*(2.0d*_dsqr[ivar] - 3.0d);
}
}
for (ivar = 0; ivar < _network.InputCount; ivar++)
{
temp = truedist*_dsqr[ivar];
_v[vsptr + ivar] += temp;
_w[wsptr + ivar] += temp
*(2.0d*_dsqr[ivar] - 3.0d);
}
psum += dist;
}
}
if (_network.OutputMode == PNNOutputMode.Classification)
{
psum = 0.0d;
for (pop = 0; pop < _network.OutputCount; pop++)
{
if (_network.Priors[pop] >= 0.0d)
{
xout[pop] *= _network.Priors[pop]
/_network.CountPer[pop];
}
psum += xout[pop];
}
if (psum < 1.0e-40d)
{
psum = 1.0e-40d;
}
}
for (pop = 0; pop < _network.OutputCount; pop++)
{
xout[pop] /= psum;
}
for (ivar = 0; ivar < _network.InputCount; ivar++)
{
if (_network.OutputMode == PNNOutputMode.Classification)
{
vtot = wtot = 0.0d;
}
else
{
vtot = _v[vsptr + ivar]*2.0d
/(psum*_network.Sigma[ivar]);
wtot = _w[wsptr + ivar]
*2.0d
/(psum*_network.Sigma[ivar]*_network.Sigma[ivar]);
}
for (outvar = 0; outvar < _network.OutputCount; outvar++)
{
if ((_network.OutputMode == PNNOutputMode.Classification)
&& (_network.Priors[outvar] >= 0.0d))
{
_v[outvar*_network.InputCount + ivar] *= _network.Priors[outvar]
/_network.CountPer[outvar];
_w[outvar*_network.InputCount + ivar] *= _network.Priors[outvar]
/_network.CountPer[outvar];
}
_v[outvar*_network.InputCount + ivar] *= 2.0d/(psum*_network.Sigma[ivar]);
_w[outvar*_network.InputCount + ivar] *= 2.0d/(psum
*_network.Sigma[ivar]*_network.Sigma[ivar]);
if (_network.OutputMode == PNNOutputMode.Classification)
{
vtot += _v[outvar*_network.InputCount + ivar];
wtot += _w[outvar*_network.InputCount + ivar];
}
}
for (outvar = 0; outvar < _network.OutputCount; outvar++)
{
der1 = _v[outvar*_network.InputCount + ivar]
- xout[outvar]*vtot;
der2 = _w[outvar*_network.InputCount + ivar]
+ 2.0d*xout[outvar]*vtot*vtot - 2.0d
*_v[outvar*_network.InputCount + ivar]
*vtot - xout[outvar]*wtot;
if (_network.OutputMode == PNNOutputMode.Classification)
{
if (outvar == (int) target[0])
{
temp = 2.0d*(xout[outvar] - 1.0d);
}
else
{
temp = 2.0d*xout[outvar];
}
}
else
{
temp = 2.0d*(xout[outvar] - target[outvar]);
}
_network.Deriv[ivar] += temp*der1;
_network.Deriv2[ivar] += temp*der2 + 2.0d*der1
*der1;
}
}
return new BasicMLData(xout);
}