/**
* Initialize weights. This includes steps for doing a random initialization of W as well as the
* vbias and hbias
*/
protected void initWeights() {
if (this.nVisible < 1)
throw new IllegalStateException("Number of visible can not be less than 1");
if (this.nHidden < 1)
throw new IllegalStateException("Number of hidden can not be less than 1");
if (this.dist == null)
dist =
new NormalDistribution(rng, 0, .01, NormalDistribution.DEFAULT_INVERSE_ABSOLUTE_ACCURACY);
/*
* Initialize based on the number of visible units..
* The lower bound is called the fan in
* The outer bound is called the fan out.
*
* Below's advice works for Denoising AutoEncoders and other
* neural networks you will use due to the same baseline guiding principles for
* both RBMs and Denoising Autoencoders.
*
* Hinton's Guide to practical RBMs:
* The weights are typically initialized to small random values chosen from a zero-mean Gaussian with
* a standard deviation of about 0.01. Using larger random values can speed the initial learning, but
* it may lead to a slightly worse final model. Care should be taken to ensure that the initial weight
* values do not allow typical visible vectors to drive the hidden unit probabilities very close to 1 or 0
* as this significantly slows the learning.
*/
if (this.W == null) {
this.W = DoubleMatrix.zeros(nVisible, nHidden);
for (int i = 0; i < this.W.rows; i++)
this.W.putRow(i, new DoubleMatrix(dist.sample(this.W.columns)));
}
this.wAdaGrad = new AdaGrad(this.W.rows, this.W.columns);
if (this.hBias == null) {
this.hBias = DoubleMatrix.zeros(nHidden);
/*
* Encourage sparsity.
* See Hinton's Practical guide to RBMs
*/
// this.hBias.subi(4);
}
this.hBiasAdaGrad = new AdaGrad(hBias.rows, hBias.columns);
if (this.vBias == null) {
if (this.input != null) {
this.vBias = DoubleMatrix.zeros(nVisible);
} else this.vBias = DoubleMatrix.zeros(nVisible);
}
this.vBiasAdaGrad = new AdaGrad(vBias.rows, vBias.columns);
}
public static void main(String argv[]) {
modshogun.init_shogun_with_defaults();
int dim = 7;
DoubleMatrix data = DoubleMatrix.rand(dim, dim);
RealFeatures feats = new RealFeatures(data);
DoubleMatrix data_T = data.transpose();
DoubleMatrix symdata = data.add(data_T);
int cols = (1 + dim) * dim / 2;
DoubleMatrix lowertriangle = DoubleMatrix.zeros(1, cols);
int count = 0;
for (int i = 0; i < dim; i++) {
for (int j = 0; j < dim; j++) {
if (j <= i) {
lowertriangle.put(0, count++, symdata.get(i, j));
}
}
}
CustomKernel kernel = new CustomKernel();
kernel.set_triangle_kernel_matrix_from_triangle(lowertriangle);
DoubleMatrix km_triangletriangle = kernel.get_kernel_matrix();
kernel.set_triangle_kernel_matrix_from_full(symdata);
DoubleMatrix km_fulltriangle = kernel.get_kernel_matrix();
kernel.set_full_kernel_matrix_from_full(data);
DoubleMatrix km_fullfull = kernel.get_kernel_matrix();
modshogun.exit_shogun();
}
/**
* Loads ground truth responses
*
* @param groundTruth is a Map from questions to responses
*/
public void setGroundTruth(Map<TypeQ, TypeR> groundTruth) {
log.debug("Setting ground truth...");
if (!(categToInt != null)) log.error("Result Vector not computed.");
assert categToInt != null : "Result Vector not computed.";
DoubleMatrix indicator = DoubleMatrix.zeros(sortedQuestions.size());
DoubleMatrix gtVec = DoubleMatrix.zeros(sortedQuestions.size());
int idx = 0;
for (TypeQ iter : sortedQuestions) {
if (groundTruth.containsKey(iter)) {
indicator.put(idx, 1.0d);
gtVec.put(idx, categToInt.get(groundTruth.get(iter)));
}
idx++;
}
groundTruthVector = new Pair<DoubleMatrix, DoubleMatrix>(gtVec, indicator);
}
/**
* Loads tune responses
*
* @param tuneSet is a Map from questions to responses
*/
public void setTune(Map<TypeQ, TypeR> tuneSet) {
log.debug("Setting tune set...");
if (!(categToInt != null)) log.error("Result Vector not computed.");
assert categToInt != null : "Result Vector not computed.";
DoubleMatrix indicator = DoubleMatrix.zeros(sortedQuestions.size());
DoubleMatrix tuneVec = DoubleMatrix.zeros(sortedQuestions.size());
int idx = 0;
for (TypeQ iter : sortedQuestions) {
if (tuneSet.containsKey(iter)) {
indicator.put(idx, 1.0d);
tuneVec.put(idx, categToInt.get(tuneSet.get(iter)));
}
idx++;
}
tuneVector = new Pair<DoubleMatrix, DoubleMatrix>(tuneVec, indicator);
}
public ClassifierTheta(double[] t, int FeatureLength, int CatSize) {
CatSize--;
Theta = t.clone();
this.FeatureLength = FeatureLength;
this.CatSize = CatSize;
W = DoubleMatrix.zeros(FeatureLength, CatSize);
b = DoubleMatrix.zeros(CatSize, 1);
if (Theta.length != CatSize * (1 + FeatureLength))
System.err.println(
"ClassifierTheta : Size Mismatch : "
+ Theta.length
+ " != "
+ (CatSize * (1 + FeatureLength)));
build();
}
@Override
public NeuralNetwork transpose() {
try {
Constructor<?> c = Dl4jReflection.getEmptyConstructor(getClass());
c.setAccessible(true);
NeuralNetwork ret = (NeuralNetwork) c.newInstance();
ret.setMomentumAfter(momentumAfter);
ret.setConcatBiases(concatBiases);
ret.setResetAdaGradIterations(resetAdaGradIterations);
ret.setVBiasAdaGrad(hBiasAdaGrad);
ret.sethBias(vBias.dup());
ret.setvBias(DoubleMatrix.zeros(hBias.rows, hBias.columns));
ret.setnHidden(getnVisible());
ret.setnVisible(getnHidden());
ret.setW(W.transpose());
ret.setL2(l2);
ret.setMomentum(momentum);
ret.setRenderEpochs(getRenderIterations());
ret.setSparsity(sparsity);
ret.setRng(getRng());
ret.setDist(getDist());
ret.setAdaGrad(wAdaGrad);
ret.setLossFunction(lossFunction);
ret.setOptimizationAlgorithm(optimizationAlgo);
ret.setConstrainGradientToUnitNorm(constrainGradientToUnitNorm);
return ret;
} catch (Exception e) {
throw new RuntimeException(e);
}
}
@Override
public void reduce(
Text key, Iterable<jBLASArrayWritable> inputs, WeightContributions.Context context)
throws IOException, InterruptedException {
DoubleMatrix w_cont = new DoubleMatrix(),
hb_cont = new DoubleMatrix(),
vb_cont = new DoubleMatrix(),
weights = null,
hbias = null,
vbias = null;
ArrayList<DoubleMatrix> chainList = new ArrayList<DoubleMatrix>();
ArrayList<DoubleMatrix> output_array = new ArrayList<DoubleMatrix>();
int count = 0;
for (jBLASArrayWritable input : inputs) {
ArrayList<DoubleMatrix> data = input.getData();
w_cont.copy(data.get(0));
hb_cont.copy(data.get(1));
vb_cont.copy(data.get(3));
// save list of all hidden chains for updates to batch files in phase 3
chainList.add(new DoubleMatrix(data.get(2).toArray2()));
if (weights == null) {
weights = DoubleMatrix.zeros(w_cont.rows, w_cont.columns);
hbias = DoubleMatrix.zeros(hb_cont.rows, hb_cont.columns);
vbias = DoubleMatrix.zeros(vb_cont.rows, vb_cont.columns);
}
// sum weight contributions
weights.addi(w_cont);
hbias.addi(hb_cont);
vbias.addi(vb_cont);
count++;
}
output_array.add(weights.div(count));
output_array.add(hbias.div(count));
output_array.add(vbias.div(count));
output_array.addAll(chainList);
jBLASArrayWritable outputmatrix = new jBLASArrayWritable(output_array);
context.write(key, outputmatrix);
}
/**
* 求逆矩阵
*
* @param mt
* @return
*/
public DoubleMatrix inverse(DoubleMatrix mt) {
int m = mt.rows;
DoubleMatrix E = DoubleMatrix.zeros(m, m);
for (int i = 0; i < m; i++) {
E.put(i, i, 1);
}
return Solve.solve(mt, E);
}
public void reduce(
IntWritable sameNum,
Iterator<Text> data,
OutputCollector<Text, jBLASArrayWritable> output,
Reporter reporter)
throws IOException {
int totalBatchCount = exampleCount / batchSize;
DoubleMatrix weights = DoubleMatrix.randn(hiddenNodes, visibleNodes);
DoubleMatrix hbias = DoubleMatrix.zeros(hiddenNodes);
DoubleMatrix vbias = DoubleMatrix.zeros(visibleNodes);
DoubleMatrix label = DoubleMatrix.zeros(1);
DoubleMatrix hidden_chain = null;
DoubleMatrix vdata = DoubleMatrix.zeros(batchSize, visibleNodes);
ArrayList<DoubleMatrix> outputmatricies = new ArrayList<DoubleMatrix>();
outputmatricies.add(weights);
outputmatricies.add(hbias);
outputmatricies.add(vbias);
outputmatricies.add(label);
outputmatricies.add(vdata);
outputmatricies.add(hidden_chain);
int j;
for (int i = 0; i < totalBatchCount; i++) {
j = 0;
while (data.hasNext() && j < batchSize) {
j++;
StringTokenizer tk = new StringTokenizer(data.next().toString());
label.put(0, Double.parseDouble(tk.nextToken()));
String image = tk.nextToken();
for (int k = 0; k < image.length(); k++) {
Integer val = new Integer(image.charAt(k));
vdata.put(j, k, val.doubleValue());
}
dataArray = new jBLASArrayWritable(outputmatricies);
batchID.set("1\t" + i);
output.collect(batchID, dataArray);
}
}
}
/** Computes results as a vector of type DoubleMatrix */
public void computeComparableResultVector() {
log.info("Computing comparable result vector...");
if (!(sortedRePrPairs != null))
log.error("Attempted to compute results vector before results.");
assert sortedRePrPairs != null : "Attempted to compute results vector before results.";
categToInt = new HashMap<TypeR, Integer>();
resultVector = DoubleMatrix.zeros(sortedQuestions.size());
categoriesMat = DoubleMatrix.zeros(responseCategories.size());
int temp = 1;
for (TypeR iter : responseCategories) {
categToInt.put(iter, temp);
categoriesMat.put(temp - 1, temp);
temp++;
}
if (log.isDebugEnabled()) log.debug("Computed map from categories to integers:\n" + categToInt);
temp = 0;
for (Pair<TypeR, Double> iter : sortedRePrPairs) {
resultVector.put(temp, categToInt.get(iter.getFirst()));
temp++;
}
if (log.isDebugEnabled()) log.debug("Computed results vector:\n" + resultVector);
log.info("Computed result vector.");
}
private void costantiniUnwrap() throws LPException {
final int ny = wrappedPhase.rows - 1; // start from Zero!
final int nx = wrappedPhase.columns - 1; // start from Zero!
if (wrappedPhase.isVector()) throw new IllegalArgumentException("Input must be 2D array");
if (wrappedPhase.rows < 2 || wrappedPhase.columns < 2)
throw new IllegalArgumentException("Size of input must be larger than 2");
// Default weight
DoubleMatrix w1 = DoubleMatrix.ones(ny + 1, 1);
w1.put(0, 0.5);
w1.put(w1.length - 1, 0.5);
DoubleMatrix w2 = DoubleMatrix.ones(1, nx + 1);
w2.put(0, 0.5);
w2.put(w2.length - 1, 0.5);
DoubleMatrix weight = w1.mmul(w2);
DoubleMatrix i, j, I_J, IP1_J, I_JP1;
DoubleMatrix Psi1, Psi2;
DoubleMatrix[] ROWS;
// Compute partial derivative Psi1, eqt (1,3)
i = intRangeDoubleMatrix(0, ny - 1);
j = intRangeDoubleMatrix(0, nx);
ROWS = grid2D(i, j);
I_J = JblasUtils.sub2ind(wrappedPhase.rows, ROWS[0], ROWS[1]);
IP1_J = JblasUtils.sub2ind(wrappedPhase.rows, ROWS[0].add(1), ROWS[1]);
Psi1 =
JblasUtils.getMatrixFromIdx(wrappedPhase, IP1_J)
.sub(JblasUtils.getMatrixFromIdx(wrappedPhase, I_J));
Psi1 = UnwrapUtils.wrapDoubleMatrix(Psi1);
// Compute partial derivative Psi2, eqt (2,4)
i = intRangeDoubleMatrix(0, ny);
j = intRangeDoubleMatrix(0, nx - 1);
ROWS = grid2D(i, j);
I_J = JblasUtils.sub2ind(wrappedPhase.rows, ROWS[0], ROWS[1]);
I_JP1 = JblasUtils.sub2ind(wrappedPhase.rows, ROWS[0], ROWS[1].add(1));
Psi2 =
JblasUtils.getMatrixFromIdx(wrappedPhase, I_JP1)
.sub(JblasUtils.getMatrixFromIdx(wrappedPhase, I_J));
Psi2 = UnwrapUtils.wrapDoubleMatrix(Psi2);
// Compute beq
DoubleMatrix beq = DoubleMatrix.zeros(ny, nx);
i = intRangeDoubleMatrix(0, ny - 1);
j = intRangeDoubleMatrix(0, nx - 1);
ROWS = grid2D(i, j);
I_J = JblasUtils.sub2ind(Psi1.rows, ROWS[0], ROWS[1]);
I_JP1 = JblasUtils.sub2ind(Psi1.rows, ROWS[0], ROWS[1].add(1));
beq.addi(JblasUtils.getMatrixFromIdx(Psi1, I_JP1).sub(JblasUtils.getMatrixFromIdx(Psi1, I_J)));
I_J = JblasUtils.sub2ind(Psi2.rows, ROWS[0], ROWS[1]);
I_JP1 = JblasUtils.sub2ind(Psi2.rows, ROWS[0].add(1), ROWS[1]);
beq.subi(JblasUtils.getMatrixFromIdx(Psi2, I_JP1).sub(JblasUtils.getMatrixFromIdx(Psi2, I_J)));
beq.muli(-1 / (2 * Constants._PI));
for (int k = 0; k < beq.length; k++) {
beq.put(k, Math.round(beq.get(k)));
}
beq.reshape(beq.length, 1);
logger.debug("Constraint matrix");
i = intRangeDoubleMatrix(0, ny - 1);
j = intRangeDoubleMatrix(0, nx - 1);
ROWS = grid2D(i, j);
DoubleMatrix ROW_I_J = JblasUtils.sub2ind(i.length, ROWS[0], ROWS[1]);
int nS0 = nx * ny;
// Use by S1p, S1m
DoubleMatrix[] COLS;
COLS = grid2D(i, j);
DoubleMatrix COL_IJ_1 = JblasUtils.sub2ind(i.length, COLS[0], COLS[1]);
COLS = grid2D(i, j.add(1));
DoubleMatrix COL_I_JP1 = JblasUtils.sub2ind(i.length, COLS[0], COLS[1]);
int nS1 = (nx + 1) * (ny);
// SOAPBinding.Use by S2p, S2m
COLS = grid2D(i, j);
DoubleMatrix COL_IJ_2 = JblasUtils.sub2ind(i.length + 1, COLS[0], COLS[1]);
COLS = grid2D(i.add(1), j);
DoubleMatrix COL_IP1_J = JblasUtils.sub2ind(i.length + 1, COLS[0], COLS[1]);
int nS2 = nx * (ny + 1);
// Equality constraint matrix (Aeq)
/*
S1p = + sparse(ROW_I_J, COL_I_JP1,1,nS0,nS1) ...
- sparse(ROW_I_J, COL_IJ_1,1,nS0,nS1);
S1m = -S1p;
S2p = - sparse(ROW_I_J, COL_IP1_J,1,nS0,nS2) ...
+ sparse(ROW_I_J, COL_IJ_2,1,nS0,nS2);
S2m = -S2p;
*/
// ToDo: Aeq matrix should be sparse from it's initialization, look into JblasMatrix factory for
// howto
// ...otherwise even a data set of eg 40x40 pixels will exhaust heap:
// ... dimension of Aeq (equality constraints) matrix for 30x30 input is 1521x6240 matrix
// ... dimension of Aeq ( ) matrix for 50x50 input is 2401x9800
// ... dimension of Aeq ( ) matrix for 512x512 input is 261121x1046528
DoubleMatrix S1p =
JblasUtils.setUpMatrixFromIdx(nS0, nS1, ROW_I_J, COL_I_JP1)
.sub(JblasUtils.setUpMatrixFromIdx(nS0, nS1, ROW_I_J, COL_IJ_1));
DoubleMatrix S1m = S1p.neg();
DoubleMatrix S2p =
JblasUtils.setUpMatrixFromIdx(nS0, nS2, ROW_I_J, COL_IP1_J)
.neg()
.add(JblasUtils.setUpMatrixFromIdx(nS0, nS2, ROW_I_J, COL_IJ_2));
DoubleMatrix S2m = S2p.neg();
DoubleMatrix Aeq =
concatHorizontally(concatHorizontally(S1p, S1m), concatHorizontally(S2p, S2m));
final int nObs = Aeq.columns;
final int nUnkn = Aeq.rows;
DoubleMatrix c1 = JblasUtils.getMatrixFromRange(0, ny, 0, weight.columns, weight);
DoubleMatrix c2 = JblasUtils.getMatrixFromRange(0, weight.rows, 0, nx, weight);
c1.reshape(c1.length, 1);
c2.reshape(c2.length, 1);
DoubleMatrix cost =
DoubleMatrix.concatVertically(
DoubleMatrix.concatVertically(c1, c1), DoubleMatrix.concatVertically(c2, c2));
logger.debug("Minimum network flow resolution");
StopWatch clockLP = new StopWatch();
LinearProgram lp = new LinearProgram(cost.data);
lp.setMinProblem(true);
boolean[] integerBool = new boolean[nObs];
double[] lowerBound = new double[nObs];
double[] upperBound = new double[nObs];
for (int k = 0; k < nUnkn; k++) {
lp.addConstraint(new LinearEqualsConstraint(Aeq.getRow(k).toArray(), beq.get(k), "cost"));
}
for (int k = 0; k < nObs; k++) {
integerBool[k] = true;
lowerBound[k] = 0;
upperBound[k] = 99999;
}
// setup bounds and integer nature
lp.setIsinteger(integerBool);
lp.setUpperbound(upperBound);
lp.setLowerbound(lowerBound);
LinearProgramSolver solver = SolverFactory.newDefault();
// double[] solution;
// solution = solver.solve(lp);
DoubleMatrix solution = new DoubleMatrix(solver.solve(lp));
clockLP.stop();
logger.debug("Total GLPK time: {} [sec]", (double) (clockLP.getElapsedTime()) / 1000);
// Displatch the LP solution
int offset;
int[] idx1p = JblasUtils.intRangeIntArray(0, nS1 - 1);
DoubleMatrix x1p = solution.get(idx1p);
x1p.reshape(ny, nx + 1);
offset = idx1p[nS1 - 1] + 1;
int[] idx1m = JblasUtils.intRangeIntArray(offset, offset + nS1 - 1);
DoubleMatrix x1m = solution.get(idx1m);
x1m.reshape(ny, nx + 1);
offset = idx1m[idx1m.length - 1] + 1;
int[] idx2p = JblasUtils.intRangeIntArray(offset, offset + nS2 - 1);
DoubleMatrix x2p = solution.get(idx2p);
x2p.reshape(ny + 1, nx);
offset = idx2p[idx2p.length - 1] + 1;
int[] idx2m = JblasUtils.intRangeIntArray(offset, offset + nS2 - 1);
DoubleMatrix x2m = solution.get(idx2m);
x2m.reshape(ny + 1, nx);
// Compute the derivative jumps, eqt (20,21)
DoubleMatrix k1 = x1p.sub(x1m);
DoubleMatrix k2 = x2p.sub(x2m);
// (?) Round to integer solution
if (roundK == true) {
for (int idx = 0; idx < k1.length; idx++) {
k1.put(idx, FastMath.floor(k1.get(idx)));
}
for (int idx = 0; idx < k2.length; idx++) {
k2.put(idx, FastMath.floor(k2.get(idx)));
}
}
// Sum the jumps with the wrapped partial derivatives, eqt (10,11)
k1.reshape(ny, nx + 1);
k2.reshape(ny + 1, nx);
k1.addi(Psi1.div(Constants._TWO_PI));
k2.addi(Psi2.div(Constants._TWO_PI));
// Integrate the partial derivatives, eqt (6)
// cumsum() method in JblasTester -> see cumsum_demo() in JblasTester.cumsum_demo()
DoubleMatrix k2_temp = DoubleMatrix.concatHorizontally(DoubleMatrix.zeros(1), k2.getRow(0));
k2_temp = JblasUtils.cumsum(k2_temp, 1);
DoubleMatrix k = DoubleMatrix.concatVertically(k2_temp, k1);
k = JblasUtils.cumsum(k, 1);
// Unwrap - final solution
unwrappedPhase = k.mul(Constants._TWO_PI);
}
public static void exportResult(
TrueSolutionCTM trueSolutionCTM,
TrueSolution[] trueSolution,
TrueSolution[] trueSolution1,
TrueSolution[] trueSolution2,
int limite,
String folder) {
try {
int bdry = 1; // *set boundary type
int numSections = trueSolution[0].numSections;
int numSections1 = trueSolution1[0].numSections;
int numSections2 = trueSolution2[0].numSections;
BufferedWriter[] writerSection = new BufferedWriter[numSections];
BufferedWriter[] writerSection1 = new BufferedWriter[numSections1];
BufferedWriter[] writerSection2 = new BufferedWriter[numSections2];
BufferedWriter[] LyapSection = new BufferedWriter[numSections];
BufferedWriter writerTrue;
BufferedWriter writerConsensusTerm;
BufferedWriter writerAvr;
BufferedWriter writerAvr1;
BufferedWriter writerAvr2;
BufferedWriter Error;
BufferedWriter Error1;
BufferedWriter Error2;
BufferedWriter Lyap;
BufferedWriter[] Mode = new BufferedWriter[numSections];
BufferedWriter[] Mode1 = new BufferedWriter[numSections1];
BufferedWriter[] Mode2 = new BufferedWriter[numSections2];
BufferedWriter[] GammaWrite = new BufferedWriter[numSections];
BufferedWriter writerDisagreement;
BufferedWriter writerDisagreement1;
BufferedWriter writerDisagreement2;
System.out.print("Beginning of simulation ");
new File("results/" + folder).mkdir();
RoadModel[] roadModels = new RoadModel[numSections];
RoadModel[] roadModels1 = new RoadModel[numSections1];
RoadModel[] roadModels2 = new RoadModel[numSections2];
for (int i = 0; i < numSections; i++) {
trueSolution[i].section = trueSolution[i].setSections();
}
for (int i = 0; i < numSections1; i++) {
trueSolution1[i].section = trueSolution1[i].setSections();
}
for (int i = 0; i < numSections2; i++) {
trueSolution2[i].section = trueSolution2[i].setSections();
}
for (int i = 0; i < numSections; i++) {
roadModels[i] = trueSolution[i].setRoadModels();
}
for (int i = 0; i < numSections1; i++) {
roadModels1[i] = trueSolution1[i].setRoadModels();
}
for (int i = 0; i < numSections2; i++) {
roadModels2[i] = trueSolution2[i].setRoadModels();
}
writerTrue =
new BufferedWriter(new FileWriter(new File("results/" + folder + "/" + "trueState.csv")));
writerConsensusTerm =
new BufferedWriter(
new FileWriter(new File("results/" + folder + "/" + "consensusTerm.csv")));
writerAvr =
new BufferedWriter(new FileWriter(new File("results/" + folder + "/" + "writerAvr.csv")));
writerAvr1 =
new BufferedWriter(
new FileWriter(new File("results/" + folder + "/" + "writerAvr1.csv")));
writerAvr2 =
new BufferedWriter(
new FileWriter(new File("results/" + folder + "/" + "writerAvr2.csv")));
Error = new BufferedWriter(new FileWriter(new File("results/" + folder + "/" + "Error.csv")));
Error1 =
new BufferedWriter(new FileWriter(new File("results/" + folder + "/" + "Error1.csv")));
Error2 =
new BufferedWriter(new FileWriter(new File("results/" + folder + "/" + "Error2.csv")));
Lyap =
new BufferedWriter(new FileWriter(new File("results/" + folder + "/" + "Lyapunov.csv")));
writerDisagreement =
new BufferedWriter(
new FileWriter(new File("results/" + folder + "/" + "Disagreement.csv")));
writerDisagreement1 =
new BufferedWriter(
new FileWriter(new File("results/" + folder + "/" + "Disagreement1.csv")));
writerDisagreement2 =
new BufferedWriter(
new FileWriter(new File("results/" + folder + "/" + "Disagreement2.csv")));
Estimation[] estimations = new Estimation[numSections];
Estimation[] estimations1 = new Estimation[numSections1];
Estimation[] estimations2 = new Estimation[numSections2];
for (int i = 0; i < numSections; i++) {
estimations[i] =
new Estimation(
roadModels[i],
"KCF",
true); // *set the last variable to indicate if consensus is needed
String S = "results/" + folder + "/" + roadModels[i].nameModel;
writerSection[i] = new BufferedWriter(new FileWriter(new File(S + ".csv")));
Mode[i] = new BufferedWriter(new FileWriter(new File(S + "Mode.csv")));
GammaWrite[i] = new BufferedWriter(new FileWriter(new File(S + "Gamma.csv")));
LyapSection[i] = new BufferedWriter(new FileWriter(new File(S + "LyapSection.csv")));
}
for (int i = 0; i < numSections1; i++) {
estimations1[i] =
new Estimation(
roadModels1[i],
"KCF",
false); // *set the last variable to indicate if consensus is needed
String S = "results/" + folder + "/" + roadModels1[i].nameModel;
writerSection1[i] = new BufferedWriter(new FileWriter(new File(S + "_1.csv")));
Mode1[i] = new BufferedWriter(new FileWriter(new File(S + "Mode1.csv")));
}
for (int i = 0; i < numSections2; i++) {
estimations2[i] =
new Estimation(
roadModels2[i],
"KCF",
false); // *set the last variable to indicate if consensus is needed
String S = "results/" + folder + "/" + roadModels2[i].nameModel;
writerSection2[i] = new BufferedWriter(new FileWriter(new File(S + "_2.csv")));
Mode2[i] = new BufferedWriter(new FileWriter(new File(S + "Mode2.csv")));
}
int overlap = roadModels[0].trueSolution.overlapsec;
int cellsec = roadModels[0].section.cellsec;
int overlap1 = roadModels1[0].trueSolution.overlapsec;
int cellsec1 = roadModels1[0].section.cellsec;
int overlap2 = roadModels2[0].trueSolution.overlapsec;
int cellsec2 = roadModels2[0].section.cellsec;
DoubleMatrix I1 = DoubleMatrix.zeros(roadModels[0].section.cellsec, overlap);
for (int i = 0; i < overlap; i++) {
I1.put(i, i, 1);
}
DoubleMatrix I2 = DoubleMatrix.zeros(roadModels[0].section.cellsec, overlap);
for (int i = roadModels[0].section.cellsec - overlap;
i < roadModels[0].section.cellsec;
i++) {
I2.put(i, i - (roadModels[0].section.cellsec - overlap), 1);
}
DoubleMatrix I123 = DoubleMatrix.concatVertically(I1.transpose(), I2.transpose());
DoubleMatrix H_hat = I2.transpose();
for (int i = 0; i < numSections - 2; i++) {
H_hat = Concat.Diagonal(H_hat, I123);
}
H_hat = Concat.Diagonal(H_hat, I1.transpose());
DoubleMatrix L_tilde =
DoubleMatrix.zeros((numSections - 1) * 2 * overlap, numSections * cellsec);
for (int i = 0; i < numSections - 1; i++) {
for (int j = 0; j < overlap; j++) {
L_tilde.put(i * (2 * overlap) + j, (i + 1) * (cellsec) + j, 1);
}
for (int j = 0; j < overlap; j++) {
L_tilde.put(i * (2 * overlap) + overlap + j, (i + 1) * (cellsec) - overlap + j, 1);
}
for (int j = 0; j < overlap; j++) {
L_tilde.put(i * (2 * overlap) + j, (i + 1) * (cellsec) - overlap + j, -1);
}
for (int j = 0; j < overlap; j++) {
L_tilde.put(i * (2 * overlap) + overlap + j, (i + 1) * (cellsec) + j, -1);
}
}
DoubleMatrix[] S = new DoubleMatrix[numSections];
trueSolutionCTM.getMeasurement();
for (int i = 0; i < numSections; i++) {
if (i == 0) {
estimations[i].filter.Ltilde = L_tilde.getRange(i, i + overlap, i, i + 2 * cellsec);
} else if (i == numSections - 1) {
estimations[i].filter.Ltilde =
L_tilde.getRange(
(2 * (numSections - 1) - 1) * overlap,
(2 * (numSections - 1) - 1) * overlap + overlap,
(i - 1) * cellsec,
(i + 1) * cellsec);
} else {
estimations[i].filter.Ltilde =
L_tilde.getRange(
(2 * i - 1) * overlap,
(2 * i + 1) * overlap,
(i - 1) * cellsec,
(i + 2) * cellsec);
}
estimations[i].filter.I1 = I1;
estimations[i].filter.I2 = I2;
}
double[] disagAvg = new double[3];
double[] errorAvg = new double[3];
double[] errAvgEntire = new double[3];
for (int k = 0; k < limite; k++) {
System.out.println("time step=" + k);
for (int i = 0; i < trueSolutionCTM.trueStatesCTM.getRows(); i++) {
writerTrue.write(trueSolutionCTM.trueStatesCTM.get(i, 0) + ",");
}
writerTrue.write("\n");
DoubleMatrix[] mean = new DoubleMatrix[numSections];
DoubleMatrix[] mean_1 = new DoubleMatrix[numSections1];
DoubleMatrix[] mean_2 = new DoubleMatrix[numSections2];
double ErrorAvg = 0;
double ErrorEntire = 0;
double ErrorAvg1 = 0;
double ErrorAvg2 = 0;
double LyapAvg = 0;
double[] errorAvgk = new double[3];
for (int i = 0; i < numSections; i++) {
mean[i] = estimations[i].getDensityMean();
if (numSections == 1) {
for (int j = 0; j < mean[i].length; j++) {
writerAvr.write(mean[i].get(j) + ",");
errorAvgk[0] =
(mean[i].get(j) - trueSolutionCTM.trueStatesCTM.get(j, 0))
* (mean[i].get(j) - trueSolutionCTM.trueStatesCTM.get(j, 0))
+ errorAvgk[0];
}
} else {
if (i == 0) {
for (int j = 0; j < mean[i].length - overlap; j++) {
writerAvr.write(mean[i].get(j) + ",");
errorAvgk[0] =
(mean[i].get(j) - trueSolutionCTM.trueStatesCTM.get(j, 0))
* (mean[i].get(j) - trueSolutionCTM.trueStatesCTM.get(j, 0))
+ errorAvgk[0];
}
} else if (i > 0 && i < numSections - 1) {
for (int j = 0; j < overlap; j++) {
writerAvr.write(
(mean[i].get(j) + mean[i - 1].get(cellsec - overlap + j)) / 2 + ",");
errorAvgk[0] =
((mean[i].get(j) + mean[i - 1].get(cellsec - overlap + j)) / 2
- trueSolutionCTM.trueStatesCTM.get(
i * (mean[i].length - overlap) + j, 0))
* ((mean[i].get(j) + mean[i - 1].get(cellsec - overlap + j)) / 2
- trueSolutionCTM.trueStatesCTM.get(
i * (mean[i].length - overlap) + j, 0))
+ errorAvgk[0];
}
for (int j = 0; j < mean[i].length - 2 * overlap; j++) {
writerAvr.write(mean[i].get(overlap + j) + ",");
errorAvgk[0] =
(mean[i].get(overlap + j)
- trueSolutionCTM.trueStatesCTM.get(
i * (mean[i].length - overlap) + overlap + j, 0))
* (mean[i].get(overlap + j)
- trueSolutionCTM.trueStatesCTM.get(
i * (mean[i].length - overlap) + overlap + j, 0))
+ errorAvgk[0];
}
} else if (i == numSections - 1) {
for (int j = 0; j < overlap; j++) {
writerAvr.write(
(mean[i].get(j) + mean[i - 1].get(cellsec - overlap + j)) / 2 + ",");
errorAvgk[0] =
((mean[i].get(j) + mean[i - 1].get(cellsec - overlap + j)) / 2
- trueSolutionCTM.trueStatesCTM.get(
i * (mean[i].length - overlap) + j, 0))
* ((mean[i].get(j) + mean[i - 1].get(cellsec - overlap + j)) / 2
- trueSolutionCTM.trueStatesCTM.get(
i * (mean[i].length - overlap) + j, 0))
+ errorAvgk[0];
}
for (int j = 0; j < mean[i].length - overlap; j++) {
writerAvr.write(mean[i].get(overlap + j) + ",");
errorAvgk[0] =
(mean[i].get(overlap + j)
- trueSolutionCTM.trueStatesCTM.get(
i * (mean[i].length - overlap) + overlap + j, 0))
* (mean[i].get(overlap + j)
- trueSolutionCTM.trueStatesCTM.get(
i * (mean[i].length - overlap) + overlap + j, 0))
+ errorAvgk[0];
}
}
}
for (int j = 0; j < mean[i].length; j++) {
writerSection[i].write(mean[i].get(j) + ",");
}
double errorSection =
(((mean[i].sub(trueSolution[i].trueStates))
.transpose()
.mmul((mean[i].sub(trueSolution[i].trueStates))))
.get(0, 0))
/ ((double) cellsec);
Error.write(errorSection + ",");
ErrorAvg = ErrorAvg + errorSection / ((double) numSections);
writerSection[i].write("\n");
double lyapunovSection =
((mean[i].sub(trueSolution[i].trueStates)).transpose())
.mmul(InverseMatrix.invPoSym(estimations[i].filter.var))
.mmul(mean[i].sub(trueSolution[i].trueStates))
.get(0, 0);
double tracevar = 0;
for (int c = 0; c < mean[i].getRows(); c++) {
tracevar = tracevar + estimations[i].filter.var.get(c, c);
}
LyapSection[i].write(lyapunovSection + ",");
LyapSection[i].write(tracevar + "\n");
if (i != 0) {
LyapAvg = LyapAvg + lyapunovSection / ((double) (numSections - 1));
}
}
errAvgEntire[0] =
errorAvgk[0] / ((double) (limite * trueSolutionCTM.cells)) + errAvgEntire[0];
Error.write(ErrorAvg + ",");
errorAvg[0] = errorAvg[0] + ErrorAvg / ((double) (limite));
Lyap.write(LyapAvg + "\n");
for (int i = 0; i < numSections1; i++) {
mean_1[i] = estimations1[i].getDensityMean();
if (numSections1 == 1) {
for (int j = 0; j < mean_1[i].length; j++) {
writerAvr1.write(mean_1[i].get(j) + ",");
errorAvgk[1] =
(mean_1[i].get(j) - trueSolutionCTM.trueStatesCTM.get(j, 0))
* (mean_1[i].get(j) - trueSolutionCTM.trueStatesCTM.get(j, 0))
+ errorAvgk[1];
}
} else {
if (i == 0) {
for (int j = 0; j < mean_1[i].length - overlap1; j++) {
writerAvr1.write(mean_1[i].get(j) + ",");
errorAvgk[1] =
(mean_1[i].get(j) - trueSolutionCTM.trueStatesCTM.get(j, 0))
* (mean_1[i].get(j) - trueSolutionCTM.trueStatesCTM.get(j, 0))
+ errorAvgk[1];
}
} else if (i > 0 && i < numSections1 - 1) {
for (int j = 0; j < overlap1; j++) {
writerAvr1.write(
(mean_1[i].get(j) + mean_1[i - 1].get(cellsec1 - overlap1 + j)) / 2 + ",");
errorAvgk[1] =
((mean_1[i].get(j) + mean_1[i - 1].get(cellsec1 - overlap1 + j)) / 2
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_1[i].length - overlap1) + j, 0))
* ((mean_1[i].get(j) + mean_1[i - 1].get(cellsec1 - overlap1 + j)) / 2
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_1[i].length - overlap1) + j, 0))
+ errorAvgk[1];
}
for (int j = 0; j < mean_1[i].length - 2 * overlap1; j++) {
writerAvr1.write(mean_1[i].get(overlap1 + j) + ",");
errorAvgk[1] =
(mean_1[i].get(overlap1 + j)
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_1[i].length - overlap1) + overlap1 + j, 0))
* (mean_1[i].get(overlap1 + j)
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_1[i].length - overlap1) + overlap1 + j, 0))
+ errorAvgk[1];
}
} else if (i == numSections1 - 1) {
for (int j = 0; j < overlap1; j++) {
writerAvr1.write(
(mean_1[i].get(j) + mean_1[i - 1].get(cellsec1 - overlap1 + j)) / 2 + ",");
errorAvgk[1] =
((mean_1[i].get(j) + mean_1[i - 1].get(cellsec1 - overlap1 + j)) / 2
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_1[i].length - overlap1) + j, 0))
* ((mean_1[i].get(j) + mean_1[i - 1].get(cellsec1 - overlap1 + j)) / 2
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_1[i].length - overlap1) + j, 0))
+ errorAvgk[1];
}
for (int j = 0; j < mean_1[i].length - overlap1; j++) {
writerAvr1.write(mean_1[i].get(overlap1 + j) + ",");
errorAvgk[1] =
(mean_1[i].get(overlap1 + j)
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_1[i].length - overlap1) + overlap1 + j, 0))
* (mean_1[i].get(overlap1 + j)
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_1[i].length - overlap1) + overlap1 + j, 0))
+ errorAvgk[1];
}
}
}
for (int j = 0; j < mean_1[i].length; j++) {
writerSection1[i].write(mean_1[i].get(j) + ",");
}
double errorSection1 =
(((mean_1[i].sub(trueSolution1[i].trueStates).transpose())
.mmul((mean_1[i].sub(trueSolution1[i].trueStates))))
.get(0, 0))
/ ((double) cellsec1);
Error1.write(errorSection1 + ",");
ErrorAvg1 = ErrorAvg1 + errorSection1 / ((double) numSections1);
writerSection1[i].write("\n");
}
Error1.write(ErrorAvg1 + ",");
errorAvg[1] = errorAvg[1] + ErrorAvg1 / ((double) (limite));
errAvgEntire[1] =
errorAvgk[1] / ((double) (limite * trueSolutionCTM.cells)) + errAvgEntire[1];
for (int i = 0; i < numSections2; i++) {
mean_2[i] = estimations2[i].getDensityMean();
if (numSections2 == 1) {
for (int j = 0; j < mean_2[i].length; j++) {
writerAvr2.write(mean_2[i].get(j) + ",");
errorAvgk[2] =
(mean_2[i].get(j) - trueSolutionCTM.trueStatesCTM.get(j, 0))
* (mean_2[i].get(j) - trueSolutionCTM.trueStatesCTM.get(j, 0))
+ errorAvgk[2];
}
} else {
if (i == 0) {
for (int j = 0; j < mean_2[i].length - overlap2; j++) {
writerAvr2.write(mean_2[i].get(j) + ",");
errorAvgk[2] =
(mean_2[i].get(j) - trueSolutionCTM.trueStatesCTM.get(j, 0))
* (mean_2[i].get(j) - trueSolutionCTM.trueStatesCTM.get(j, 0))
+ errorAvgk[2];
}
} else if (i > 0 && i < numSections2 - 1) {
for (int j = 0; j < overlap2; j++) {
writerAvr2.write(
(mean_2[i].get(j) + mean_2[i - 1].get(cellsec2 - overlap2 + j)) / 2 + ",");
errorAvgk[2] =
((mean_2[i].get(j) + mean_2[i - 1].get(cellsec2 - overlap2 + j)) / 2
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_2[i].length - overlap2) + j, 0))
* ((mean_2[i].get(j) + mean_2[i - 1].get(cellsec2 - overlap2 + j)) / 2
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_2[i].length - overlap2) + j, 0))
+ errorAvgk[2];
}
for (int j = 0; j < mean_2[i].length - 2 * overlap2; j++) {
writerAvr2.write(mean_2[i].get(overlap2 + j) + ",");
errorAvgk[2] =
(mean_2[i].get(overlap2 + j)
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_2[i].length - overlap2) + overlap2 + j, 0))
* (mean_2[i].get(overlap2 + j)
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_2[i].length - overlap2) + overlap2 + j, 0))
+ errorAvgk[2];
}
} else if (i == numSections2 - 1) {
for (int j = 0; j < overlap2; j++) {
writerAvr2.write(
(mean_2[i].get(j) + mean_2[i - 1].get(cellsec2 - overlap2 + j)) / 2 + ",");
errorAvgk[2] =
((mean_2[i].get(j) + mean_2[i - 1].get(cellsec2 - overlap2 + j)) / 2
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_2[i].length - overlap2) + j, 0))
* ((mean_2[i].get(j) + mean_2[i - 1].get(cellsec2 - overlap2 + j)) / 2
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_2[i].length - overlap2) + j, 0))
+ errorAvgk[2];
}
for (int j = 0; j < mean_2[i].length - overlap2; j++) {
writerAvr2.write(mean_2[i].get(overlap2 + j) + ",");
errorAvgk[2] =
(mean_2[i].get(overlap2 + j)
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_2[i].length - overlap2) + overlap2 + j, 0))
* (mean_2[i].get(overlap2 + j)
- trueSolutionCTM.trueStatesCTM.get(
i * (mean_2[i].length - overlap2) + overlap2 + j, 0))
+ errorAvgk[2];
}
}
}
for (int j = 0; j < mean_2[i].length; j++) {
writerSection2[i].write(mean_2[i].get(j) + ",");
}
double errorSection2 =
(((mean_2[i].sub(trueSolution2[i].trueStates))
.transpose()
.mmul((mean_2[i].sub(trueSolution2[i].trueStates))))
.get(0, 0))
/ ((double) (cellsec2));
Error2.write(errorSection2 + ",");
ErrorAvg2 = ErrorAvg2 + errorSection2 / ((double) numSections2);
writerSection2[i].write("\n");
}
Error2.write(ErrorAvg2 + ",");
errorAvg[2] = errorAvg[2] + ErrorAvg2 / ((double) (limite));
errAvgEntire[2] =
errorAvgk[2] / ((double) (limite * trueSolutionCTM.cells)) + errAvgEntire[2];
writerAvr.write("\n");
writerAvr1.write("\n");
writerAvr2.write("\n");
Error.write("\n");
Error1.write("\n");
Error2.write("\n");
Section[] secs = new Section[numSections];
Section[] secs1 = new Section[numSections1];
Section[] secs2 = new Section[numSections2];
for (int i = 0; i < numSections; i++) {
secs[i] = trueSolution[i].setSections(); // every step, update mode
}
for (int i = 0; i < numSections1; i++) {
secs1[i] = trueSolution1[i].setSections();
}
for (int i = 0; i < numSections2; i++) {
secs2[i] = trueSolution2[i].setSections();
}
for (int i = 0; i < numSections; i++) {
secs[i].Estimates = mean[i];
}
for (int i = 0; i < numSections1; i++) {
secs1[i].Estimates = mean_1[i];
}
for (int i = 0; i < numSections2; i++) {
secs2[i].Estimates = mean_2[i];
}
for (int i = 0; i < numSections; i++) {
if (secs[i].getClass().getSimpleName().equals("FC")) {
secs[i].getwavefrontfromEstimation();
if (secs[i].wavefront < 0) {
secs[i].wavefront = 0;
}
if (secs[i].wavefront <= cellsec - 2) {
if (trueSolution[i].flux(secs[i].Estimates.get(secs[i].wavefront, 0))
- trueSolution[i].flux(secs[i].Estimates.get(secs[i].wavefront + 1, 0))
> 0) {
secs[i].wavedirection = 0;
} else {
secs[i].wavedirection = 1;
}
} else if (secs[i].wavefront == cellsec - 1) {
if (secs[i].index != numSections - 1) {
if (trueSolution[i].flux(secs[i].Estimates.get(secs[i].wavefront, 0))
- trueSolution[i + 1].flux(secs[i].Estimates.get(overlap, 0))
> 0) {
secs[i].wavedirection = 0;
} else {
secs[i].wavedirection = 1;
}
} else {
secs[i].wavedirection = 0;
}
}
secs[i].getModelA();
secs[i].getModelB1();
secs[i].getModelB2();
} else if (secs[i].getClass().getSimpleName().equals("CF")) {
secs[i].getwavefrontfromEstimation();
secs[i].getModelA();
secs[i].getModelB1();
secs[i].getModelB2();
}
}
for (int i = 0; i < numSections1; i++) {
if (secs1[i].getClass().getSimpleName().equals("FC")) {
secs1[i].getwavefrontfromEstimation();
if (secs1[i].wavefront < 0) {
secs1[i].wavefront = 0;
}
if (secs1[i].wavefront <= cellsec1 - 2) {
if (trueSolution1[i].flux(secs1[i].Estimates.get(secs1[i].wavefront, 0))
- trueSolution1[i].flux(secs1[i].Estimates.get(secs1[i].wavefront + 1, 0))
> 0) {
secs1[i].wavedirection = 0;
} else {
secs1[i].wavedirection = 1;
}
} else if (secs1[i].wavefront == cellsec1 - 1) {
if (secs1[i].index != numSections1 - 1) {
if (trueSolution1[i].flux(secs1[i].Estimates.get(secs1[i].wavefront, 0))
- trueSolution1[i + 1].flux(secs1[i].Estimates.get(overlap1, 0))
> 0) {
secs1[i].wavedirection = 0;
} else {
secs1[i].wavedirection = 1;
}
} else {
secs1[i].wavedirection = 0;
}
}
secs1[i].getModelA();
secs1[i].getModelB1();
secs1[i].getModelB2();
} else if (secs1[i].getClass().getSimpleName().equals("CF")) {
secs1[i].getwavefrontfromEstimation();
secs1[i].getModelA();
secs1[i].getModelB1();
secs1[i].getModelB2();
}
}
for (int i = 0; i < numSections2; i++) {
if (secs2[i].getClass().getSimpleName().equals("FC")) {
secs2[i].getwavefrontfromEstimation();
if (secs2[i].wavefront < 0) {
secs2[i].wavefront = 0;
}
if (secs2[i].wavefront <= cellsec2 - 2) {
if (trueSolution2[i].flux(secs2[i].Estimates.get(secs2[i].wavefront, 0))
- trueSolution2[i].flux(secs2[i].Estimates.get(secs2[i].wavefront + 1, 0))
> 0) {
secs2[i].wavedirection = 0;
} else {
secs2[i].wavedirection = 1;
}
} else if (secs2[i].wavefront == cellsec2 - 1) {
if (secs2[i].index != numSections2 - 1) {
if (trueSolution2[i].flux(secs2[i].Estimates.get(secs2[i].wavefront, 0))
- trueSolution2[i + 1].flux(secs2[i].Estimates.get(overlap2, 0))
> 0) {
secs2[i].wavedirection = 0;
} else {
secs2[i].wavedirection = 1;
}
} else {
secs2[i].wavedirection = 0;
}
}
secs2[i].getModelA();
secs2[i].getModelB1();
secs2[i].getModelB2();
} else if (secs2[i].getClass().getSimpleName().equals("CF")) {
secs2[i].getwavefrontfromEstimation();
secs2[i].getModelA();
secs2[i].getModelB1();
secs2[i].getModelB2();
}
}
for (int i = 0; i < numSections; i++) {
estimations[i].setSection(secs[i]);
}
for (int i = 0; i < numSections1; i++) {
estimations1[i].setSection(secs1[i]);
}
for (int i = 0; i < numSections2; i++) {
estimations2[i].setSection(secs2[i]);
}
for (int i = 0; i < numSections; i++) {
trueSolution[i].section = secs[i];
}
for (int i = 0; i < numSections1; i++) {
trueSolution1[i].section = secs1[i];
}
for (int i = 0; i < numSections2; i++) {
trueSolution2[i].section = secs2[i];
}
for (int i = 0; i < numSections; i++) {
if (i == 0) {
GammaWrite[i].write(estimations[i].filter.gammaStar + ",");
} else if (i == numSections - 1) {
GammaWrite[i].write(estimations[i].filter.gammaStar + ",");
} else {
GammaWrite[i].write(estimations[i].filter.gammaStar + ",");
}
GammaWrite[i].write("\n");
}
for (int i = 0; i < numSections; i++) {
Mode[i].write(estimations[i].roadModel.section.getClass().getSimpleName() + ",");
Mode[i].write(
estimations[i].roadModel.section.wavefront + ","); // estimated location of shock
Mode[i].write(
estimations[i].roadModel.section._wavefront + ","); // true location of shock
Mode[i].write(estimations[i].roadModel.section.wavedirection + ",");
Mode[i].write(estimations[i].roadModel.section._wavedirection + ",");
Mode[i].write("\n");
}
for (int i = 0; i < numSections1; i++) {
Mode1[i].write(estimations1[i].roadModel.section.getClass().getSimpleName() + ",");
Mode1[i].write(estimations1[i].roadModel.section.wavefront + ",");
Mode1[i].write(estimations1[i].roadModel.section._wavefront + ",");
Mode1[i].write(estimations1[i].roadModel.section.wavedirection + ",");
Mode1[i].write(estimations1[i].roadModel.section._wavedirection + ",");
Mode1[i].write("\n");
}
for (int i = 0; i < numSections2; i++) {
Mode2[i].write(estimations2[i].roadModel.section.getClass().getSimpleName() + ",");
Mode2[i].write(estimations2[i].roadModel.section.wavefront + ",");
Mode2[i].write(estimations2[i].roadModel.section._wavefront + ",");
Mode2[i].write(estimations2[i].roadModel.section.wavedirection + ",");
Mode2[i].write(estimations2[i].roadModel.section._wavedirection + ",");
Mode2[i].write("\n");
}
for (int i = 0; i < numSections; i++) {
// trueSolution[i].updateMeasurement();
// roadModels[i].updateMeasurement();
estimations[i].filter.getNewParametersFromModel();
}
for (int i = 0; i < numSections1; i++) {
// trueSolution[i].updateMeasurement();
// roadModels[i].updateMeasurement();
estimations1[i].filter.getNewParametersFromModel();
}
for (int i = 0; i < numSections2; i++) {
// trueSolution[i].updateMeasurement();
// roadModels[i].updateMeasurement();
estimations2[i].filter.getNewParametersFromModel();
}
for (int i = 0; i < numSections; i++) {
S[i] =
(estimations[i].filter.measure.transpose())
.mmul(InverseMatrix.invPoSym(estimations[i].filter.measureVar))
.mmul(estimations[i].filter.measure);
}
double disag = 0;
for (int i = 0; i < numSections - 1; i++) {
disag =
disag
+ (((mean[i]
.getRange(cellsec - overlap, cellsec, 0, 1)
.sub(mean[i + 1].getRange(0, overlap, 0, 1)))
.transpose()
.mmul(
mean[i]
.getRange(cellsec - overlap, cellsec, 0, 1)
.sub(mean[i + 1].getRange(0, overlap, 0, 1))))
.get(0, 0))
/ ((double) overlap);
}
disag = disag / ((double) (numSections - 1));
writerDisagreement.write(disag + ",");
disagAvg[0] = disagAvg[0] + disag / ((double) limite);
for (int i = 0; i < numSections - 1; i++) {
writerDisagreement.write(
((mean[i]
.getRange(cellsec - overlap, cellsec, 0, 1)
.sub(mean[i + 1].getRange(0, overlap, 0, 1)))
.transpose()
.mmul(
mean[i]
.getRange(cellsec - overlap, cellsec, 0, 1)
.sub(mean[i + 1].getRange(0, overlap, 0, 1))))
.get(0, 0)
+ ",");
}
writerDisagreement.write("\n");
double disag1 = 0;
for (int i = 0; i < numSections1 - 1; i++) {
disag1 =
disag1
+ (((mean_1[i]
.getRange(cellsec1 - overlap1, cellsec1, 0, 1)
.sub(mean_1[i + 1].getRange(0, overlap1, 0, 1)))
.transpose()
.mmul(
mean_1[i]
.getRange(cellsec1 - overlap1, cellsec1, 0, 1)
.sub(mean_1[i + 1].getRange(0, overlap1, 0, 1))))
.get(0, 0))
/ ((double) overlap1);
}
disag1 = disag1 / ((double) (numSections1 - 1));
writerDisagreement1.write(disag1 + ",");
disagAvg[1] = disagAvg[1] + disag1 / ((double) limite);
for (int i = 0; i < numSections1 - 1; i++) {
writerDisagreement1.write(
((mean_1[i]
.getRange(cellsec1 - overlap1, cellsec1, 0, 1)
.sub(mean_1[i + 1].getRange(0, overlap1, 0, 1)))
.transpose()
.mmul(
mean_1[i]
.getRange(cellsec1 - overlap1, cellsec1, 0, 1)
.sub(mean_1[i + 1].getRange(0, overlap1, 0, 1))))
.get(0, 0)
+ ",");
}
writerDisagreement1.write("\n");
double disag2 = 0;
for (int i = 0; i < numSections2 - 1; i++) {
disag2 =
disag2
+ (((mean_2[i]
.getRange(cellsec2 - overlap2, cellsec2, 0, 1)
.sub(mean_2[i + 1].getRange(0, overlap2, 0, 1)))
.transpose()
.mmul(
mean_2[i]
.getRange(cellsec2 - overlap2, cellsec2, 0, 1)
.sub(mean_2[i + 1].getRange(0, overlap2, 0, 1))))
.get(0, 0))
/ ((double) overlap2);
}
disag2 = disag2 / ((double) (numSections2 - 1));
writerDisagreement2.write(disag2 + ",");
disagAvg[2] = disagAvg[2] + disag2 / ((double) limite);
for (int i = 0; i < numSections2 - 1; i++) {
writerDisagreement2.write(
((mean_2[i]
.getRange(cellsec2 - overlap2, cellsec2, 0, 1)
.sub(mean_2[i + 1].getRange(0, overlap2, 0, 1)))
.transpose()
.mmul(
mean_2[i]
.getRange(cellsec2 - overlap2, cellsec2, 0, 1)
.sub(mean_2[i + 1].getRange(0, overlap2, 0, 1))))
.get(0, 0)
+ ",");
}
writerDisagreement2.write("\n");
trueSolutionCTM.update();
// **start defining boundary condition
int cells = trueSolutionCTM.trueStatesCTMPrior.getRows();
if (bdry == 0) {
double inflow = trueSolutionCTM.receiving(trueSolutionCTM.trueStatesCTMPrior.get(0, 0));
double outflow =
trueSolutionCTM.sending(trueSolutionCTM.trueStatesCTMPrior.get(cells - 1, 0));
double upOut =
trueSolutionCTM.computeFlux(
trueSolutionCTM.trueStatesCTMPrior.get(0, 0),
trueSolutionCTM.trueStatesCTMPrior.get(1, 0));
trueSolutionCTM.trueStatesCTM.put(
0,
0,
trueSolutionCTM.trueStatesCTMPrior.get(0, 0)
+ (trueSolutionCTM.dt / trueSolutionCTM.dx) * (inflow - upOut));
double downIn =
trueSolutionCTM.computeFlux(
trueSolutionCTM.trueStatesCTMPrior.get(cells - 2, 0),
trueSolutionCTM.trueStatesCTMPrior.get(cells - 1, 0));
trueSolutionCTM.trueStatesCTM.put(
cells - 1,
0,
trueSolutionCTM.trueStatesCTMPrior.get(cells - 1, 0)
+ (trueSolutionCTM.dt / trueSolutionCTM.dx) * (downIn - outflow));
} else if (bdry == 1) {
double inflow = 0.1125 + 0.1125 * Math.sin(k * (Math.PI / 4000) + (Math.PI));
if (inflow > trueSolutionCTM.receiving(trueSolutionCTM.trueStatesCTMPrior.get(0, 0))) {
inflow = trueSolutionCTM.receiving(trueSolutionCTM.trueStatesCTMPrior.get(0, 0));
}
double outflow =
trueSolutionCTM.receiving(trueSolutionCTM.trueStatesCTMPrior.get(cells - 1, 0));
// double outflow=0.1125+0.1125*Math.sin(k*(Math.PI/1000));
// if
// (outflow>trueSolution[numSections-1].sending(trueSolution[numSections-1].trueStatesPrior.get(cellsec-1,0))){
//
// outflow=trueSolution[numSections-1].sending(trueSolution[numSections-1].trueStatesPrior.get(cellsec-1,0));
// }
double upOut =
trueSolutionCTM.computeFlux(
trueSolutionCTM.trueStatesCTMPrior.get(0, 0),
trueSolutionCTM.trueStatesCTMPrior.get(1, 0));
trueSolutionCTM.trueStatesCTM.put(
0,
0,
trueSolutionCTM.trueStatesCTMPrior.get(0, 0)
+ (trueSolutionCTM.dt / trueSolutionCTM.dx) * (inflow - upOut));
double downIn =
trueSolutionCTM.computeFlux(
trueSolutionCTM.trueStatesCTMPrior.get(cells - 2, 0),
trueSolutionCTM.trueStatesCTMPrior.get(cells - 1, 0));
trueSolutionCTM.trueStatesCTM.put(
cells - 1,
0,
trueSolutionCTM.trueStatesCTMPrior.get(cells - 1, 0)
+ (trueSolutionCTM.dt / trueSolutionCTM.dx) * (downIn - outflow));
} else {
double inflow = 0.1125 + 0.1125 * Math.sin(k * (Math.PI / 8000) + (Math.PI));
if (inflow > trueSolutionCTM.receiving(trueSolutionCTM.trueStatesCTMPrior.get(0, 0))) {
inflow = trueSolutionCTM.receiving(trueSolutionCTM.trueStatesCTMPrior.get(0, 0));
}
double outflow =
trueSolutionCTM.receiving(trueSolutionCTM.trueStatesCTMPrior.get(cells - 1, 0));
// double outflow=0.1125+0.1125*Math.sin(k*(Math.PI/1000));
// if
// (outflow>trueSolution[numSections-1].sending(trueSolution[numSections-1].trueStatesPrior.get(cellsec-1,0))){
//
// outflow=trueSolution[numSections-1].sending(trueSolution[numSections-1].trueStatesPrior.get(cellsec-1,0));
// }
double upOut =
trueSolutionCTM.computeFlux(
trueSolutionCTM.trueStatesCTMPrior.get(0, 0),
trueSolutionCTM.trueStatesCTMPrior.get(1, 0));
trueSolutionCTM.trueStatesCTM.put(
0,
0,
trueSolutionCTM.trueStatesCTMPrior.get(0, 0)
+ (trueSolutionCTM.dt / trueSolutionCTM.dx) * (inflow - upOut));
double downIn =
trueSolutionCTM.computeFlux(
trueSolutionCTM.trueStatesCTMPrior.get(cells - 2, 0),
trueSolutionCTM.trueStatesCTMPrior.get(cells - 1, 0));
trueSolutionCTM.trueStatesCTM.put(
cells - 1,
0,
trueSolutionCTM.trueStatesCTMPrior.get(cells - 1, 0)
+ (trueSolutionCTM.dt / trueSolutionCTM.dx) * (downIn - outflow));
// *define boundary conditions here
}
// **end defining boundary condition
trueSolutionCTM.getMeasurement();
for (int i = 0; i < numSections; i++) {
trueSolution[i].update();
}
for (int i = 0; i < numSections; i++) {
trueSolution[i].updatemeasurement();
}
for (int i = 0; i < numSections1; i++) {
trueSolution1[i].update();
}
for (int i = 0; i < numSections1; i++) {
trueSolution1[i].updatemeasurement();
}
for (int i = 0; i < numSections2; i++) {
trueSolution2[i].update();
}
for (int i = 0; i < numSections2; i++) {
trueSolution2[i].updatemeasurement();
}
for (int i = 0; i < numSections; i++) {
estimations[i].nextStepWithoutUpdateTrue();
}
for (int i = 0; i < numSections1; i++) {
estimations1[i].nextStepWithoutUpdateTrue();
}
for (int i = 0; i < numSections2; i++) {
estimations2[i].nextStepWithoutUpdateTrue();
}
if (estimations[0].consensus) {
for (int i = 0; i < numSections; i++) {
estimations[i].filter.ComputeEig();
}
for (int i = 0; i < numSections; i++) {
if (i == 0) {
estimations[i].filter.ComputeGammaStar2(estimations[1].filter);
} else if (i == numSections - 1) {
estimations[i].filter.ComputeGammaStar1(estimations[numSections - 2].filter);
} else {
estimations[i].filter.ComputeGammaStar(
estimations[i - 1].filter, estimations[i + 1].filter);
}
}
DoubleMatrix[] mean1 = new DoubleMatrix[numSections];
mean1[0] = estimations[0].filter.Consensus2(estimations[1].filter);
writerConsensusTerm.write(estimations[0].filter.scaling.get(0) + ",");
writerConsensusTerm.write(estimations[0].filter.scaling.get(1) + ",");
for (int i = 1; i < numSections - 1; i++) {
mean1[i] =
estimations[i].filter.Consensus(
estimations[i - 1].filter, estimations[i + 1].filter);
writerConsensusTerm.write(estimations[i].filter.scaling.get(0) + ",");
writerConsensusTerm.write(estimations[i].filter.scaling.get(1) + ",");
}
mean1[numSections - 1] =
estimations[numSections - 1].filter.Consensus1(estimations[numSections - 2].filter);
writerConsensusTerm.write(estimations[numSections - 1].filter.scaling.get(0) + ",");
writerConsensusTerm.write(estimations[numSections - 1].filter.scaling.get(1) + ",");
writerConsensusTerm.write("\n");
for (int i = 0; i < numSections; i++) {
estimations[i].filter.mean = mean1[i];
}
}
}
for (int i = 0; i < numSections; i++) {
writerSection[i].flush();
writerSection[i].close();
Mode[i].flush();
Mode[i].close();
LyapSection[i].flush();
LyapSection[i].close();
GammaWrite[i].flush();
GammaWrite[i].close();
}
for (int i = 0; i < numSections1; i++) {
writerSection1[i].flush();
writerSection1[i].close();
Mode1[i].flush();
Mode1[i].close();
}
for (int i = 0; i < numSections2; i++) {
writerSection2[i].flush();
writerSection2[i].close();
Mode2[i].flush();
Mode2[i].close();
}
writerAvr.flush();
writerAvr.close();
writerConsensusTerm.flush();
writerConsensusTerm.close();
writerAvr1.flush();
writerAvr1.close();
writerAvr2.flush();
writerAvr2.close();
Error.flush();
Error.close();
Error1.flush();
Error1.close();
Error2.flush();
Error2.close();
Lyap.flush();
Lyap.close();
writerTrue.flush();
writerTrue.close();
writerDisagreement.flush();
writerDisagreement.close();
writerDisagreement1.flush();
writerDisagreement1.close();
writerDisagreement2.flush();
writerDisagreement2.close();
System.out.print("Disag_avg_DLKCF=" + disagAvg[0] + "\n");
System.out.print("Disag_avg_DLKF=" + disagAvg[1] + "\n");
System.out.print("Disag_avg_LKF=" + disagAvg[2] + "\n");
System.out.print("Err_avg_DLKCF=" + errorAvg[0] + "\n");
System.out.print("Err_avg_DLKF=" + errorAvg[1] + "\n");
System.out.print("Err_avg_LKF=" + errorAvg[2] + "\n");
System.out.println(" End");
} catch (Exception e) {
e.printStackTrace();
}
}
public double train(DataSet train_data) {
_n_inputs = train_data._n_cols;
int n_class = train_data._n_classes;
_n_outputs = Math.min(_n_inputs, n_class - 1);
build();
// train_data.print_dat_format();
DoubleMatrix Sw = DoubleMatrix.zeros(_n_inputs, _n_inputs);
DoubleMatrix Sb = DoubleMatrix.zeros(_n_inputs, _n_inputs);
int[] n_classes = new int[n_class];
double[][] mean_for_class = new double[n_class][_n_inputs];
double[] mean = new double[_n_inputs];
int n_examples = train_data._n_rows, i, j;
for (i = 0; i < n_examples; i++) {
double[] X = train_data.get_X(i);
int label = train_data.get_label(i);
for (j = 0; j < _n_inputs; j++) {
mean_for_class[label][j] += X[j];
mean[j] += X[j];
}
n_classes[label]++;
}
double[][] diff_class_mean = new double[n_class][_n_inputs];
for (j = 0; j < _n_inputs; j++) {
mean[j] /= 1.0 * n_examples; // compute the total mean value;
for (i = 0; i < n_class; i++) {
if (n_classes[i] == 0) {
mean_for_class[i][j] = 0;
} else {
mean_for_class[i][j] /= 1.0 * n_classes[i];
}
diff_class_mean[i][j] = mean_for_class[i][j] - mean[j];
}
}
// OutFile.printf(mean_for_class);
for (i = 0; i < n_class; i++) {
double[] diff = diff_class_mean[i];
// Sb = Sb + c_i (v_i - mean)(v_i - mean);
Sb.addi(new DoubleMatrix(Vec.outer_dot(diff, diff)).muli(n_classes[i]));
}
// for(i = 0; i < _n_inputs; i++)
// {
// for(j = 0; j < _n_inputs; j++)
// {
// OutFile.printf("%f ",Sb.get(i,j));
// }
// OutFile.printf("\n");
// }
for (i = 0; i < n_examples; i++) {
double[] diff = Vec.minus(train_data.get_X(i), mean_for_class[train_data.get_label(i)]);
// Sw = Sw + diff * diff
Sw.addi(new DoubleMatrix(Vec.outer_dot(diff, diff)));
}
DoubleMatrix[] eig_sw = Eigen.symmetricEigenvectors(Sw);
// for(i = 0; i < _n_inputs; i++)
// {
// for(j = 0; j < _n_inputs; j++)
// {
// OutFile.printf("%f ",eig_sw[0].get(i,j));
// }
// OutFile.printf("\n");
// }
double s;
for (j = 0; j < _n_inputs; j++) {
s = eig_sw[1].get(j, j);
// OutFile.printf("%f\n",s);
if (s > General.SMALL_CONST) {
s = 1.0 / Math.sqrt(s);
} else {
s = 0;
}
for (i = 0; i < _n_inputs; i++) {
eig_sw[0].put(i, j, eig_sw[0].get(i, j) * s);
}
}
// eig_sw[0].print();
// OutFile.printf(" sf.\n");
// for(i = 0; i < _n_inputs; i++)
// {
// for(j = 0; j < _n_inputs; j++)
// {
// OutFile.printf("%f ",eig_sw[0].get(i,j));
// }
// OutFile.printf("\n");
// }
// eig_sw[0].print();
Sb = eig_sw[0].transpose().mmul(Sb).mmul(eig_sw[0]);
// OutFile.printf(" sb.\n");
// for(i = 0; i < _n_inputs; i++)
// {
// for(j = 0; j < _n_inputs; j++)
// {
// OutFile.printf("%f ",Sb.get(i,j));
// }
// OutFile.printf("\n");
// }
DoubleMatrix[] eig_sb = Eigen.symmetricEigenvectors(Sb);
double sum = 0;
for (i = 0; i < _n_inputs; i++) {
sum += eig_sb[1].get(i, i);
}
OutFile.printf("eig value: \n");
for (i = 0; i < _n_inputs; i++) {
OutFile.printf("%f ", eig_sb[1].get(i, i) / sum);
}
OutFile.printf("\n");
eig_sw[0].mmuli(eig_sb[0]);
// //eig_vec[0].print();
// for(i = 0; i < _n_inputs; i++)
// {
// for(j = 0; j < _n_inputs; j++)
// {
// OutFile.printf("%f ",eig_sb[1].get(i,j));
// }
// OutFile.printf("\n");
// }
// OutFile.printf("%d %d outputs: %d\n",eig_vec[0].rows,eig_vec[0].columns, _n_outputs);
for (i = 0; i < _n_outputs; i++) {
for (j = 0; j < _n_inputs; j++) {
_eig_mat[j][i] = eig_sw[0].get(j, _n_inputs - 1 - i);
}
}
return 0.0f;
}
public static void main(String[] args) {
Scanner s = new Scanner(System.in);
int n = s.nextInt();
int nn = n;
DoubleMatrix d = DoubleMatrix.zeros(2 * n, 2 * n);
DoubleMatrix d2 = DoubleMatrix.zeros(n, n);
for (int i = 0; i < n; i++) {
for (int j = 0; j < n; j++) {
d.put(i, j, s.nextDouble());
d2.put(i, j, d.get(i, j));
}
}
// d2 = new DoubleMatrix(d.data);
System.out.println(d);
System.out.println(d);
List<Integer> L = new ArrayList<Integer>(2 * n);
List<Double> R = new ArrayList<Double>(2 * n);
for (int i = 0; i < 2 * n; i++) {
L.add(0);
R.add(0.0);
}
// inicjalizacja lisci - jako etykiety kolejne liczby od 0
for (int i = 0; i < n; i++) {
L.set(i, i);
}
// V - drzewo addytywne, które tworzymy
ArrayList[] V = new ArrayList[2 * n];
for (int i = 0; i < V.length; i++) {
V[i] = new ArrayList<Integer>();
}
double suma, rmin, rr;
int i, j, vertNum = n;
while (n > 3) {
// wyznaczanie r dla każdego liścia
for (int a = 0; a < n; a++) {
suma = 0;
for (int b = 0; b < n; b++) {
suma = suma + d.get(L.get(a), L.get(b));
}
suma = suma / (n - 2);
R.set(a, suma);
}
// wyznaczania sąsiadów na podstawie r
i = 0;
j = 1;
rmin = d.get(L.get(0), L.get(1)) - (R.get(0) + R.get(1));
for (int a = 0; a < n - 1; a++) {
for (int b = a + 1; b < n; b++) {
rr = d.get(L.get(a), L.get(b)) - (R.get(a) + R.get(b));
if (rr < rmin) {
rmin = rr;
i = a;
j = b;
}
}
}
// usuniecie ze zbioru lisci i,j oraz dodanie k
L.set(n, vertNum);
vertNum++;
i = L.remove(i);
j = L.remove(j - 1);
n = n - 1;
// uaktualnienie d dla każdego pozostałego liścia
for (int l = 0; l < n - 1; l++) {
double value = (d.get(i, L.get(l)) + d.get(j, L.get(l)) - d.get(i, j)) / 2;
d.put(L.get(n - 1), L.get(l), value);
d.put(L.get(l), L.get(n - 1), value);
}
// dodanie odpowiednich krawędzi do tworzonego drzewa
V[i].add((vertNum - 1));
V[j].add((vertNum - 1));
V[vertNum - 1].add(i);
V[vertNum - 1].add(j);
// wyznaczenie odległości między nowym wierzchołkiem oraz i,j
double value = (d.get(i, j) + d.get(i, L.get(0)) - d.get(j, L.get(0))) / 2;
d.put(i, vertNum - 1, value);
d.put(vertNum - 1, i, value);
d.put(j, vertNum - 1, d.get(i, j) - d.get(i, vertNum - 1));
d.put(vertNum - 1, j, d.get(i, j) - d.get(i, vertNum - 1));
}
// 3 elementowe drzewo
double value;
value = (d.get(L.get(0), L.get(1)) + d.get(L.get(0), L.get(2)) - d.get(L.get(1), L.get(2))) / 2;
d.put(L.get(0), vertNum, value);
d.put(vertNum, L.get(0), value);
value = (d.get(L.get(0), L.get(1)) + d.get(L.get(1), L.get(2)) - d.get(L.get(0), L.get(2))) / 2;
d.put(L.get(1), vertNum, value);
d.put(vertNum, L.get(1), value);
value = (d.get(L.get(0), L.get(2)) + d.get(L.get(1), L.get(2)) - d.get(L.get(0), L.get(1))) / 2;
d.put(L.get(2), vertNum, value);
d.put(vertNum, L.get(2), value);
V[vertNum].add(L.get(0));
V[vertNum].add(L.get(1));
V[vertNum].add(L.get(2));
V[L.get(0)].add(vertNum);
V[L.get(1)].add(vertNum);
V[L.get(2)].add(vertNum);
// wypisanie wyników
System.out.println(d);
// DoubleMatrix w2 = DoubleMatrix.zeros(2*n, 2*n);
ArrayList w = new ArrayList<Integer>();
for (int a = 0; a <= vertNum; a++) {
System.out.print(a);
System.out.print(" : ");
for (int b = 0; b < V[a].size(); b++) {
System.out.print(V[a].get(b));
System.out.print(" ");
// w2.put(a,b,Integer.parseInt(V[a].get(b).toString()));
w.add(V[a].get(b));
}
System.out.println("");
}
DoubleMatrix A = DoubleMatrix.zeros((nn * (nn - 1)) / 2, vertNum);
DoubleMatrix g = DoubleMatrix.zeros((nn * (nn - 1)) / 2, 1);
double blad = nk(A, d2, g, V, vertNum); // wrzucam to do siebie - mkd
System.out.println(A);
DoubleMatrix p = (new DoubleMatrix(A.rows, A.columns, A.data)).transpose();
System.out.println(p.rows + " " + p.columns);
DoubleMatrix p2 =
(new DoubleMatrix(p.rows, p.columns, p.data))
.mmul((new DoubleMatrix(A.rows, A.columns, A.data)));
System.out.println("p2: " + p2);
DoubleMatrix p3 = MatrixFunctions.pow(p2, -1);
System.out.println("p3: " + p3);
DoubleMatrix p4 = p3.mmul(p);
DoubleMatrix b = p4.mmul(g);
// DoubleMatrix b = MatrixFunctions.pow(A.transpose().mmul(A), -1).mmul(A.transpose()).mmul(g);
System.out.println(g);
System.out.println(b);
System.out.println("Kwadrat bledu wynosi " + blad);
}
public void propagateGround() {
DoubleMatrix _densitynext = DoubleMatrix.zeros(trueStatesGround.getRows(), 1);
trueStatesGroundPrior = trueStatesGround.dup();
_densitynext = AMatrix.mmul(trueStatesGroundPrior);
trueStatesGround = _densitynext.dup();
}
// TODO: DOUBLECHECK EVERYTHING
@Override
public void map(Text key, jBLASArrayWritable input, GibbsSampling.Context context)
throws IOException, InterruptedException {
/* *******************************************************************/
/* initialize all memory we're going to use during the process */
long start_time = System.nanoTime();
ArrayList<DoubleMatrix> data = input.getData();
label = data.get(4);
v_data = data.get(5);
// check to see if we are in the first layer or there are layers beneath us we must sample
// from
if (data.size() > 6) {
int prelayer = (data.size() - 6) / 3;
DoubleMatrix[] preWeights = new DoubleMatrix[prelayer],
preHbias = new DoubleMatrix[prelayer],
preVbias = new DoubleMatrix[prelayer];
for (int i = 0; i < prelayer; i++) {
preWeights[i] = data.get(6 + i * 3);
preHbias[i] = data.get(7 + i * 3);
preVbias[i] = data.get(8 + i * 3);
}
DoubleMatrix vnew = null;
for (int i = 0; i < prelayer; i++) {
weights = preWeights[i];
vbias = preVbias[i];
hbias = preHbias[i];
vnew = sample_h_from_v(i == 0 ? v_data : vnew);
}
v_data = vnew;
}
weights = data.get(0);
hbias = data.get(1);
hiddenChain = data.get(2);
vbias = data.get(3);
// check if we need to attach labels to the observed variables
if (vbias.columns != v_data.columns) {
DoubleMatrix labels = DoubleMatrix.zeros(1, classCount);
int labelNum = (new Double(label.get(0))).intValue();
labels.put(labelNum, 1.0);
v_data = DoubleMatrix.concatHorizontally(v_data, labels);
}
w1 = DoubleMatrix.zeros(weights.rows, weights.columns);
hb1 = DoubleMatrix.zeros(hbias.rows, hbias.columns);
vb1 = DoubleMatrix.zeros(vbias.rows, vbias.columns);
/* ********************************************************************/
// sample hidden state to get positive phase
// if empty, use it as the start of the chain
// or use persistent hidden state from pCD
DoubleMatrix phaseSample = sample_h_from_v(v_data);
h1_data = new DoubleMatrix();
v1_data = new DoubleMatrix();
if (hiddenChain == null) {
data.set(2, new DoubleMatrix(hbias.rows, hbias.columns));
hiddenChain = data.get(2);
hiddenChain.copy(phaseSample);
h1_data.copy(phaseSample);
} else {
h1_data.copy(hiddenChain);
}
// run Gibbs chain for k steps
for (int j = 0; j < gibbsSteps; j++) {
v1_data.copy(sample_v_from_h(h1_data));
h1_data.copy(sample_h_from_v(v1_data));
}
DoubleMatrix hprob = propup(v1_data);
weight_contribution(hiddenChain, v_data, hprob, v1_data);
hiddenChain.copy(h1_data);
data.get(0).copy(w1);
data.get(1).copy(hb1);
data.get(2).copy(hiddenChain);
data.get(3).copy(vb1);
jBLASArrayWritable outputmatrix = new jBLASArrayWritable(data);
context.write(key, outputmatrix);
log.info("Job completed in: " + (System.nanoTime() - start_time) / (1E6) + " ms");
}
