private static Vector <float> PointFToVector(PointF p)
{
float[] xy = new float[2];
xy[0] = p.X;
xy[1] = p.Y;
return(CreateVector.DenseOfArray(xy));
}
public double[,] SortRank(double[,] orgRanks)
{
Matrix <double> matRanks = CreateMatrix.DenseOfArray(orgRanks);
double[] projects = matRanks.Column(0).ToArray();
double[] ranks = matRanks.Column(1).ToArray();
Array.Sort(ranks, projects, Comparer <double> .Create((a, b) =>
{
if (a == 0 && b != 0)
{
return(1);
}
if (a != 0 && b == 0)
{
return(-1);
}
if (a > b)
{
return(1);
}
if (a < b)
{
return(-1);
}
return(0);
}));
matRanks.SetColumn(0, CreateVector.DenseOfArray(projects));
matRanks.SetColumn(1, CreateVector.DenseOfArray(ranks));
return(matRanks.ToArray());
}
/// <summary>
/// Runs through the network and makes a prediction.
/// </summary>
/// <param name="input">The input image to run the prediction on.</param>
/// <returns>The values from the final layer. The largest value is the networks prediction.</returns>
public double[] FeedForward(double[] input)
{
Matrix <double> _currentImages = CreateMatrix.DenseOfRowVectors(CreateVector.DenseOfArray <double>(input));
foreach (ILayerInformation _layerInformation in this.LayerInformation)
{
switch (_layerInformation.LayerType)
{
case (LayerType.Convolutional):
ConvolutionalLayerInformation _convInfo = _layerInformation as ConvolutionalLayerInformation;
_currentImages = this.Convolve(_convInfo, _currentImages);
break;
case (LayerType.Pooling):
PoolingLayerInformation _poolInfo = _layerInformation as PoolingLayerInformation;
_currentImages = this.Pool(_poolInfo, _currentImages);
break;
case (LayerType.NonLinear):
NonLinearLayerInformation _nonLinearInfo = _layerInformation as NonLinearLayerInformation;
_currentImages = this.NonLinear(_nonLinearInfo, _currentImages);
break;
}
}
double[] _fullyConnectedInput = CreateVector.DenseOfEnumerable <double>(_currentImages.EnumerateRows().SelectMany(a => a)).ToArray();
return(this.FullyConnectedNetwork.FeedForward(_currentImages.Enumerate().ToArray()));
}
//Called after gathering training data
public Vector <double>[] Solve(double[,] X, double[,] yArray)
{
//Solve matrix to generate weights for runtime calibration
double[] y = new double[yArray.GetLength(0)];
Vector <double>[] theta = new Vector <double> [yArray.GetLength(1)];
//Process each collum one by one
for (int i = 0; i < yArray.GetLength(1); i++)
{
//Clone raw data to manipulate
double[,] Xclone = X.Clone() as double[, ];
//Disable channels not used for this solve
Xclone = DisableChannels(Xclone, ActiveSensors, i);
Matrix <double> XMatrix = CreateMatrix.DenseOfArray(Xclone);
RemoveChannels temp = RemoveEmptyChannels(XMatrix);
XMatrix = temp.ReducedMatrix;
//Get a solution vector
y = getCollumn(yArray, i);
Matrix <double> tempMat = XMatrix;
//Perform Regression fit
Vector <double> yVector = CreateVector.DenseOfArray(y);
theta[i] = MultipleRegression.Svd <double>(tempMat, yVector);
//Add removed collumns
theta[i] = PadArray(temp.usedCol, theta[i]);
}
return(theta);
}
public void TrainingSet(double[,] x, int[] y)
{
if (x.GetUpperBound(0) + 1 < y.Length)
{
throw new Exception("You have more labels than training examples!");
}
if (x.GetUpperBound(0) + 1 > y.Length)
{
throw new Exception("You have more training examples than labels!");
}
int cv_size = (int)(0.2 * (x.GetUpperBound(0) + 1));
double[] _y = new double[y.Length - 2 * cv_size];
y = (int[])y.Clone();
X = CreateMatrix.DenseOfArray(x);
Random r = new Random();
for (int i = 0; i < X.RowCount - 2; i++)
{
int j = r.Next(i, X.RowCount);
var a = X.Row(i);
var b = X.Row(j);
X.SetRow(j, a);
X.SetRow(i, b);
int t = y[i];
y[i] = y[j];
y[j] = t;
}
CV = CreateMatrix.DenseOfRowVectors(X.Row(0));
X = X.RemoveRow(0);
Test = CreateMatrix.DenseOfRowVectors(X.Row(0));
X = X.RemoveRow(0);
for (int i = 0; i < cv_size - 1; i++)
{
CV = CV.InsertRow(i + 1, X.Row(0));
X = X.RemoveRow(0);
Test = Test.InsertRow(i + 1, X.Row(0));
X = X.RemoveRow(0);
}
double[] _cvy = new double[cv_size];
double[] _testy = new double[cv_size];
for (int i = 0; i < 2 * cv_size; i++)
{
_cvy[i / 2] = y[i++];
_testy[i / 2] = y[i];
}
for (int i = 2 * cv_size; i < y.Length; i++)
{
_y[i - 2 * cv_size] = y[i];
}
Y = CreateVector.DenseOfArray(_y);
CVY = CreateVector.DenseOfArray(_cvy);
TestY = CreateVector.DenseOfArray(_testy);
}
public int Predict(double[] x)
{
if (x.Length != Layers[0])
{
throw new Exception("The input layer requires " + Layers[0] + " values. You gave " + x.Length + ".");
}
return(Predict(CreateVector.DenseOfArray(x)));
}
Vector <double> CurrentStateVector()
{
float angularRotation = GetAngularRotationInRadians();
double[] newStateArray = { _rb.position.x, _rb.velocity.x,
_rb.position.y, _rb.velocity.y,
angularRotation, _rb.angularVelocity };
return(CreateVector.DenseOfArray <double>(newStateArray));
}
public static Vector <double> Normalize(byte[] data)
{
double[] ret = new double[data.Length];
for (int i = 0; i < ret.Length; i++)
{
ret[i] = data[i] / 255.0;
}
return(CreateVector.DenseOfArray <double>(ret));
}
public VectorBentimark()
{
vec1 = Enumerable.Range(0, size).Select(x => (float)x).ToArray();
vec2 = Enumerable.Range(0, size).Select(x => (float)x).ToArray();
this.v1 = CreateVector.DenseOfArray(vec1);
this.v2 = CreateVector.DenseOfArray(vec2);
this.mv1 = new Vector(vec1);
this.mv2 = new Vector(vec2);
}
public void feedForward(double[] pattern)
{
Neurons[0] = CreateVector.DenseOfArray(pattern);
for (int l = 0, ln = Biases.Length; l < ln; l++)
{
Neurons[l + 1] = (Weights[l].LeftMultiply(Neurons[l])).Add(Biases[l]);
Neurons[l + 1].MapInplace(Activations[l]);
}
}
public VerifyResult VerifyRank(double[,] initRanks, double[,] expAwards)
{
// init
Matrix <double> matInitRanks = CreateMatrix.DenseOfArray(initRanks).Clone(); // make a copy so don't effect the original ranks
Matrix <double> matExpAwards = CreateMatrix.DenseOfArray(expAwards);
// clean up matRanks to filter out the rows have rank of 0
Tuple <int, double> zero = matInitRanks.Column(1).Find(r => r == 0.0d, Zeros.Include);
if (zero != null && zero.Item1 > 0)
{
matInitRanks = matInitRanks.SubMatrix(0, zero.Item1, 0, matInitRanks.ColumnCount);
}
Vector <double> vecActRankNumber = matInitRanks.Column(1);
Vector <double> vecExpAwards = CreateVector.DenseOfArray(To1DAwards(expAwards));
// init dimension
int row = Math.Max(vecActRankNumber.Count, vecExpAwards.Count);
int column = matExpAwards.RowCount;
// matrix of Actual
Matrix <double> matActRankNo = ToRankMatrix(row, column, vecActRankNumber.ToArray());
// matrix of Expected
Matrix <double> matExpAwardsNo = ToRankMatrix(row, column, vecExpAwards.ToArray());
// diff between the matrix of Actual and the matrix of Expected
Matrix <double> matDiff = matExpAwardsNo.Subtract(matActRankNo);
if (matDiff.ForAll(d => d == 0.0d) /*.Equals(CreateMatrix.Dense(row, column, 0.0d)*/)
{
return(new VerifyResult()
{
Pass = true,
OptRanks = matInitRanks.ToArray(),
OrgRanks = initRanks
});
}
else
{
double[,] actAwards = To2DAwards(vecActRankNumber.ToArray());
return(new VerifyResult()
{
Pass = false,
DiffMatrix = matDiff.ToArray(),
ActRankNoMatrix = matActRankNo.ToArray(),
ActAwards = actAwards,
ExpAwards = expAwards,
OptRanks = matInitRanks.ToArray(),
OrgRanks = initRanks
});
}
}
public NonlinearMinimizationResult FindMinimum(IObjectiveModel objective, double[] initialGuess,
double[] lowerBound = null, double[] upperBound = null, double[] scales = null, bool[] isFixed = null)
{
var lb = (lowerBound == null) ? null : CreateVector.Dense <double>(lowerBound);
var ub = (upperBound == null) ? null : CreateVector.Dense <double>(upperBound);
var sc = (scales == null) ? null : CreateVector.Dense <double>(scales);
var fx = (isFixed == null) ? null : isFixed.ToList();
return(Minimum(Subproblem, objective, CreateVector.DenseOfArray <double>(initialGuess), lb, ub, sc, fx,
GradientTolerance, StepTolerance, FunctionTolerance, RadiusTolerance, MaximumIterations));
}
public VectorBentimark(int size)
{
this.size = size;
Random r = new Random(0);
vec1 = Enumerable.Range(0, size).Select(x => (float)r.Next()).ToArray();
vec2 = Enumerable.Range(0, size).Select(x => (float)r.Next()).ToArray();
this.v1 = CreateVector.DenseOfArray(vec1);
this.v2 = CreateVector.DenseOfArray(vec2);
this.mv1 = new Vector(vec1);
this.mv2 = new Vector(vec2);
}
static void Main(string[] args)
{
var matrix = CreateMatrix.DenseOfArray <double>(new double[, ] {
{ 1.0, -1.0, 2.0 },
{ 0.0, -3.0, 1.0 }
});
var vector = CreateVector.Dense <double>(new double[] { 2, 1, 0 });
//Matrix-Vector product
Console.WriteLine(MatrixDotVector(matrix, vector)); //Expected: [1 -3]
//Hadamard Procut
Console.WriteLine(HadamardProduct(vector, vector)); //Expected: [4,1,0]
//Pointwise substraction
Console.WriteLine(PointwiseSubstraction(vector, vector)); //Expected: [0,0,0]
Console.WriteLine(PointwiseSubstraction(matrix, matrix)); //Expected: [0,0,0][0,0,0]
//Mask a row/column (but not clear it)
var uMatrix = CreateMatrix.Dense <double>(2, 3, 1.0);
uMatrix.ClearColumns(0, 2);
var cMatrix = CreateMatrix.Dense <double>(2, 3, 1.0);
cMatrix.ClearColumn(1);
//with 1st and 3rd columns masked
var newMatrix = matrix.PointwiseMultiply(uMatrix);
Console.WriteLine(newMatrix);
//original matrix
Console.WriteLine(matrix);
//update the newMatrix
newMatrix[0, 1] = 100;
newMatrix[1, 1] = 200;
Console.WriteLine(newMatrix);
//push new values back to original
matrix = matrix.PointwiseMultiply(cMatrix) + newMatrix.PointwiseMultiply(uMatrix);
Console.WriteLine(matrix);
//Vector equality
Vector <double> a1 = CreateVector.DenseOfArray <double>(new double[] { 1.2, 3.4, 5.6 });
Vector <double> a2 = CreateVector.DenseOfArray <double>(new double[] { 1.2, 3.4, 5.6 });
Console.WriteLine(a1 == a2);
Console.WriteLine(a1.Equals(a2));
//Find biggest element in vector
Vector <double> c = CreateVector.DenseOfArray <double>(new double[] { 1.2, 3.4, 5.6 });
Console.WriteLine(c.MaximumIndex());
}
/// <summary>
///
/// </summary>
public void SetLocalPVectorCOG(Matrix <double> KeepMatrix)
{
int ElementIndex = GetElementIndex();
Vector <double> TestCOG = CreateVector.DenseOfArray <double>(new double[3] {
1, 2, 3
});
///this.LocalPVectorCOG = this.KeepMatrixIdentity.Transpose() * GetLocalVectorMatrix(ElementIndex, Cog.Offset); /// Offset from parent origin to local origin missing
this.LocalPVectorCOG = GetLocalVectorMatrix(ElementIndex, TestCOG);
this.LocalPVectorCOG.InsertAtIndex(IdentityVectors.X, ElementIndex);
this.LocalPVectorCOG.InsertAtIndex(IdentityVectors.Y, ElementIndex + 3);
this.LocalPVectorCOG.InsertAtIndex(IdentityVectors.Z, ElementIndex + 6);
this.LocalPVectorCOG = KeepMatrix * this.LocalPVectorCOG;
}
public void IterateKalman()
{
if (pauseWhileNotMoving && prevMousePos == curMousePos)
{
return;
}
// add mesurement spread to the mouse pos
var mesuredPos = curMousePos;
mesuredPos.Offset(rnd.NextDouble(), rnd.NextDouble());
// prepare mesurement Vector
var M = CreateVector.DenseOfArray(new[] {
mesuredPos.X,
mesuredPos.Y,
});
// calc state vector prediction
var Xkp = A * Xprev + B * a;
// calc covariance matrix prediction
var Pkp = A * (Pprev * A.Transpose()) + Q;
// calc kalman gain
var K = (Pkp * H.Transpose()) * (H * (Pkp * H.Transpose()) + R).Inverse();
// calc new state
var Xk = Xkp + K * (M - H * Xkp);
// calc new covariance matrix
var Pk = (CreateMatrix.DiagonalIdentity <double>(4) - K * H) * Pkp;
// save current state for next iteration
Xprev = Xk;
Pprev = Pk;
// update UI
OnPropertyChanged(nameof(XSpeed));
OnPropertyChanged(nameof(YSpeed));
// draw predicted line
AddToLine(inkCanvasEstimated1, new Point(Xkp[0], Xkp[1]), Colors.Blue);
// draw corrected line
AddToLine(inkCanvasEstimated2, new Point(Xprev[0], Xprev[1]), Colors.Green);
// draw mesurment points
DrawPoint(mesuredPos, Colors.Red);
prevMousePos = curMousePos;
}
private void DLSIterate()
{
var jacobian = ComputeJacobian();
var deltaE = SafeDeltaE(goal, baseJoint.EndEffector);
var jacobianMat = CreateMatrix.Dense(2, jacobian.Count, (row, col) => row == 0 ? jacobian[col].X : jacobian[col].Y);
var lambda = .5f;
var lambdaMatrix = lambda * lambda * CreateMatrix.DenseIdentity <float>(jacobianMat.RowCount);
var deltaPhiVec =
jacobianMat.Transpose() *
(jacobianMat * jacobianMat.Transpose() + (lambdaMatrix)).Inverse() *
CreateVector.DenseOfArray(new float[] { deltaE.X, deltaE.Y });
var deltaPhi = new LinkedList <float>(deltaPhiVec.ToArray());
baseJoint.ApplyDofDeltas(deltaPhi);
}
public void IterateDLS(Hinge joint, Vector2 goal)
{
var jacobian = ComputeJacobian(joint);
var deltaE = ClampDeltaE(goal, joint.EndEffector);
var jacobianMat = CreateMatrix.Dense(2, jacobian.Count, (row, col) => row == 0 ? jacobian[col].X : jacobian[col].Y);
var lambda = DLSLambda;
var lambdaMatrix = lambda * lambda * CreateMatrix.DenseIdentity <float>(jacobianMat.RowCount);
var deltaPhiVec =
jacobianMat.Transpose() *
(jacobianMat * jacobianMat.Transpose() + (lambdaMatrix)).Inverse() *
CreateVector.DenseOfArray(new float[] { deltaE.X, deltaE.Y });
var deltaPhi = new LinkedList <float>(deltaPhiVec.ToArray());
joint.ApplyDofDeltas(deltaPhi);
}
/// <summary>
/// [offset] [type] [value] [description]
///0000 32 bit integer 0x00000801(2049) magic number(MSB first)
///0004 32 bit integer 60000 number of items
///0008 unsigned byte ?? label
///0009 unsigned byte ?? label
///........
///xxxx unsigned byte ?? label
///The labels values are 0 to 9.
/// </summary>
/// <param name="labelFile"></param>
/// <returns></returns>
public static List <Vector <double> > LoadLabels(string labelFile)
{
if (!File.Exists(labelFile))
{
throw new FileNotFoundException();
}
List <byte> ret = new List <byte>();
using (FileStream stream = File.OpenRead(labelFile))
{
using (GZipStream uncompressed = new GZipStream(stream, CompressionMode.Decompress))
{
using (BinaryReader reader = new BinaryReader(uncompressed))
{
int magicNumber = toLittleEndian(reader.ReadInt32());
if (magicNumber != 0x801)
{
throw new InvalidDataException("0x801 is expected at offset 000");
}
int labelCount = toLittleEndian(reader.ReadInt32());
ret.AddRange(reader.ReadBytes(labelCount));
}
}
}
List <Vector <double> > vectors = new List <Vector <double> >();
foreach (var label in ret)
{
vectors.Add(CreateVector.DenseOfArray <double>(
new double[]
{
label == 0? 1.0:0.0,
label == 1? 1.0:0.0,
label == 2? 1.0:0.0,
label == 3? 1.0:0.0,
label == 4? 1.0:0.0,
label == 5? 1.0:0.0,
label == 6? 1.0:0.0,
label == 7? 1.0:0.0,
label == 8? 1.0:0.0,
label == 9? 1.0:0.0
}));
}
return(vectors);
}
public NonlinearMinimizationResult FindMinimum(IObjectiveModel objective, double[] initialGuess,
double[] lowerBound = null, double[] upperBound = null, double[] scales = null, bool[] isFixed = null)
{
if (objective == null)
{
throw new ArgumentNullException("objective");
}
if (initialGuess == null)
{
throw new ArgumentNullException("initialGuess");
}
var lb = (lowerBound == null) ? null : CreateVector.Dense <double>(lowerBound);
var ub = (upperBound == null) ? null : CreateVector.Dense <double>(upperBound);
var sc = (scales == null) ? null : CreateVector.Dense <double>(scales);
var fx = (isFixed == null) ? null : isFixed.ToList();
return(Minimum(objective, CreateVector.DenseOfArray <double>(initialGuess), lb, ub, sc, fx, InitialMu, GradientTolerance, StepTolerance, FunctionTolerance, MaximumIterations));
}
public static List <Vector <double> > GetOAT(Matrix <double> rot33)
{
var TOLERANCE = 1E-09;
var st2 = Math.Sqrt(rot33[0, 2] * rot33[0, 2] + rot33[1, 2] * rot33[1, 2]);
double a1, o1, t1;
double a2, o2, t2;
if (TOLERANCE < Math.Abs(st2))
{
var st = st2;
a1 = Math.Atan2(+st, rot33[2, 2]);
a2 = Math.Atan2(-st, rot33[2, 2]);
o1 = Math.Atan2(rot33[1, 2], +rot33[0, 2]);
o2 = Math.Atan2(rot33[1, 2], -rot33[0, 2]);
t1 = Math.Atan2(rot33[2, 1], -rot33[2, 0]);
t2 = Math.Atan2(rot33[2, 1], +rot33[2, 0]);
}
else
{
//o1+t1=Math.Atan2(rot33[1, 0], rot33[1, 1]) を満たす (o1, t1)の組み合わせは無数にある。
//ここでは、t1=0 という解を得ることにする。
o1 = Math.Atan2(rot33[1, 0], rot33[1, 1]);
a1 = (1d - rot33[2, 2]) * Math.PI / 2d;
t1 = 0d;
o2 = double.NaN;
a2 = double.NaN;
t2 = double.NaN;
}
var ans = new List <Vector <double> >();
ans.Add(CreateVector.DenseOfArray(new double[] { o1, a1, t1 }));
if (!double.IsNaN(a2) || !double.IsNaN(a2) || !double.IsNaN(t2))
{
ans.Add(CreateVector.DenseOfArray(new double[] { o2, a2, t2 }));
}
return(ans);
}
//Add zeroed channels to align weights with their appropriate value
private Vector <double> PadArray(bool [] usedCol, Vector <double> theta)
{
//Add removed collumns to array to make later calcs easier
int index = 0;
double[] tempArray = new double[usedCol.Length];
for (int j = 0; j < usedCol.Length; j++)
{
if (usedCol[j])
{
tempArray[j] = theta.At(index);
index++;
}
else
{
tempArray[j] = 0;
}
}
return(CreateVector.DenseOfArray(tempArray));
}
void PlotOutputs()
{
if (!HasPerceptron)
{
return;
}
// precompute
int ax = comboX.Active;
int ay = comboY.Active;
int volume = DISCRET_LEVEL * DISCRET_LEVEL;
double step = 2.0 / (DISCRET_LEVEL + 1.0);
// create points
Vector3D[][] points = new Vector3D[conf.numOutputs][];
for (int o = 0; o < conf.numOutputs; o++)
{
points[o] = new Vector3D[volume];
}
// fill points
int c = 0;
for (int ix = 1; ix <= DISCRET_LEVEL; ix++)
{
sector[ax] = step * ix;
for (int iy = 1; iy <= DISCRET_LEVEL; iy++)
{
sector[ay] = step * iy;
Vector <double> point = CreateVector.DenseOfArray(sector);
double[] values = ann.Response(point).ToArray();
for (int o = 0; o < conf.numOutputs; o++)
{
points[o][c++] = new Vector3D(sector[ax], sector[ay], values[o]);
}
}
}
// render points
rs3DP.Render(points, surf.Width, surf.Height);
}
public void MoveCar()
{
if (_car.HasCollidedWithWall)
{
_car.Move(0f, -1f);
return;
}
float speedNormalized = _car.CurrentSpeed / _car.MaxSpeed;
float[] sensorDistancesNormalized = new float[_probe.SensorCount];
_probe.SensorDistances.CopyTo(sensorDistancesNormalized, 0);
Array.ForEach(sensorDistancesNormalized, distance => distance = distance / _probe.MaxSensorDistance);
float[] networkInputs = new float[_probe.SensorCount + 1];
networkInputs[0] = speedNormalized;
sensorDistancesNormalized.CopyTo(networkInputs, 1);
Vector <float> networkOutputs = _network.FeedForward(CreateVector.DenseOfArray(networkInputs));
_car.Move(networkOutputs[0], networkOutputs[1]);
}
public static Layer Load(StreamReader reader)
{
var strContent = reader.ReadLine();
var values = strContent.Split(';');
int biascount = int.Parse(values[0]);
int colCount = int.Parse(values[1]);
int rowCount = int.Parse(values[2]);
strContent = reader.ReadLine();
values = strContent.Split(';');
var biasArray = new double[biascount];
for (int i = 0; i < biascount; i++)
{
biasArray[i] = double.Parse(values[i]);
}
var bias = CreateVector.DenseOfArray <double>(biasArray);
if (colCount != 0 && rowCount != 0)
{
double[,] mValues = new double[rowCount, colCount];
for (int y = 0; y < rowCount; y++)
{
strContent = reader.ReadLine();
values = strContent.Split(';');
for (int x = 0; x < colCount; x++)
{
mValues[y, x] = double.Parse(values[x]);
}
}
var weights = CreateMatrix.DenseOfArray(mValues);
return(new Layer(bias, weights));
}
else
{
return(new Layer(bias, null));
}
}
private OrgRank SplitRank(Matrix <double> matRanks, int rank_length, double[] awardNos, double[,] awards)
{
Vector <double> vecAwardNos = CreateVector.DenseOfArray(awardNos);
//fill the rest of ranks as the last award
double next_award = vecAwardNos[rank_length];
for (int i = rank_length; i < matRanks.RowCount; i++)
{
if (i < matRanks.RowCount)
{
matRanks[i, 1] = next_award;
}
}
OrgRank orgRank = new OrgRank(matRanks.ToArray(), awards);
// split ranks into two groups
Matrix <double> matRanks1 = matRanks.SubMatrix(0, rank_length, 0, matRanks.ColumnCount);
// split award_numbers into two sub_awards
double[,] awards1 = To2DAwards(vecAwardNos.SubVector(0, rank_length).ToArray());
Rank rank1 = new Rank(matRanks1.ToArray(), awards1);
orgRank.AddSubRank(rank1);
// split ranks into two groups
int length = matRanks.RowCount - rank_length;
Matrix <double> matRanks2 = matRanks.SubMatrix(rank_length, length, 0, matRanks.ColumnCount);
// split award_numbers into two sub_awards
length = awardNos.Length - rank_length;
double[,] awards2 = To2DAwards(vecAwardNos.SubVector(rank_length, length).ToArray());
Rank rank2 = new Rank(matRanks2.ToArray(), awards2);
orgRank.AddSubRank(rank2);
return(orgRank);
}
public vnnCm(double[][,] W, double[][] B, Func <double, double>[] activations)
{
Activations = activations;
Weights = new Matrix <double> [W.Length];
for (int i = 0; i < W.Length; i++)
{
Weights[i] = CreateMatrix.DenseOfArray(W[i]);
}
Biases = new Vector <double> [B.Length];
for (int i = 0; i < B.Length; i++)
{
Biases[i] = CreateVector.DenseOfArray(B[i]);
}
Neurons = new Vector <double> [B.Length + 1];
//Neurons[0] = CreateVector.DenseOfArray(new double[Weights[0].RowCount]);
for (int i = 1; i < Neurons.Length; i++)
{
Neurons[i] = CreateVector.DenseOfArray(new double[Biases[i - 1].Count]);
}
}
static void Main(string[] args)
{
NeuralNetwork nn = new NeuralNetwork(0.001f, 1f, 2, 2, 2);
//Console.WriteLine(nn.ToString());
double[] arr1 = { 0, 0 };
double[] arr2 = { 0, 1 };
double[] arr3 = { 1, 1 };
double[] arr4 = { 1, 0 };
double[] t1 = { 0 };
double[] t2 = { 1 };
double[] t3 = { 0 };
double[] t4 = { 1 };
Vector <double> v1 = CreateVector.DenseOfArray(arr1);
Vector <double> v2 = CreateVector.DenseOfArray(arr2);
Vector <double> v3 = CreateVector.DenseOfArray(arr3);
Vector <double> v4 = CreateVector.DenseOfArray(arr4);
Vector <double> w1 = CreateVector.DenseOfArray(t1);
Vector <double> w2 = CreateVector.DenseOfArray(t2);
Vector <double> w3 = CreateVector.DenseOfArray(t3);
Vector <double> w4 = CreateVector.DenseOfArray(t4);
Vector <double>[] input = { v1, v2, v3, v4 };
Vector <double>[] result = { w1, w2, w3, w4 };
//Console.WriteLine(v1);
//Console.WriteLine(nn.AddBias(v1));
//Console.WriteLine(nn.FeedForward(v2));
nn.Train(input, result);
//Console.WriteLine();
Console.Write("Input vector\n" + v1.ToString() + "Prediction\n" + nn.FeedForward(v1) + "\nCorrect result\n" + w1.ToString());
Console.Write("Input vector\n" + v2.ToString() + "Prediction\n" + nn.FeedForward(v2) + "\nCorrect result\n" + w2.ToString());
Console.Write("Input vector\n" + v3.ToString() + "Prediction\n" + nn.FeedForward(v3) + "\nCorrect result\n" + w3.ToString());
Console.Write("Input vector\n" + v4.ToString() + "Prediction\n" + nn.FeedForward(v4) + "\nCorrect result\n" + w4.ToString());
}
public Point SolveRayScreenVector(Point3D vec, ScreenProperties properties)
{
var unitPermm = UnitPerMM;
var tempVec = CreateVector.Dense(vec.ToArray());
var tempScale = Math.Abs(LandmarkTransformVector[2] / tempVec[2]);
tempVec = tempVec * tempScale;
tempVec = tempVec + CreateVector.DenseOfArray(LandmarkTransformVector);
if (tempVec[2] > 0.001)
{
Logger.Error("Wrong Vector try to solve");
}
var tempScr = properties.ToScreenCoordinate(unitPermm, new Point3D(tempVec.ToArray()));
//tempScr.X = properties.PixelSize.Width - Util.FixZero(tempScr.X);
//tempScr.Y = -Util.FixZero(tempScr.Y);
tempScr.X = Util.FixZero(tempScr.X);
tempScr.Y = Util.FixZero(tempScr.Y);
return(tempScr);
}
static Vector <double> makeSpock()
{
return(CreateVector.DenseOfArray <double>(new double[] { 0, 0, 0, 0, 1 }));
}
