protected void compute() { nny = (int) ni[0].parseInput(); df.writeLine("Number of events is: n = " + nny); y = DatanRandom.standardNormal(nny); out(y); mllg = new MinLogLikeGauss(y); x0 = new DatanVector(2); x0.setElement(0, 1.); x0.setElement(1, 2.); MinSim ms = new MinSim(x0, mllg); x = ms.getMinPosition(); fcont = ms.getMinimum(); df.writeLine("Minimization with MinSim yields x = "); df.writeLine(x.toString()); // covariance matrix MinCov mc = new MinCov(x, mllg); cx = mc.getCovarianceMatrix(1.); df.writeLine("Covariance Matrix cx = "); df.writeLine(cx.toString()); // asymmetric errors fcont = fcont + 0.5; MinAsy ma = new MinAsy(x, cx, mllg); dxasy = ma.getAsymmetricErrors(fcont); DatanMatrix as = new DatanMatrix(dxasy); df.writeLine("Asymmetic errors:"); df.writeLine(as.toString()); df.writeLine("ma.hasConverged() = " + ma.hasConverged()); plotScatterDiagram(); plotParameterPlane(); }
protected void compute() { n = (int) ni[0].parseInput(); t0 = ni[1].parseInput(); deltat = ni[2].parseInput(); x1 = ni[3].parseInput(); x2 = ni[4].parseInput(); sigma = ni[5].parseInput(); // generate data points t = new DatanVector(n); y = new DatanVector(n); dy = new DatanVector(n); rand = DatanRandom.standardNormal(n); for (int i = 0; i < n; i++) { t.setElement(i, t0 + (double) i * deltat); y.setElement(i, x1 * Math.pow(t.getElement(i), x2) + sigma * rand[i]); dy.setElement(i, sigma); } // find 1st approximation of unknowns by method of log-log plot tlog = new double[n]; ylog = new double[n]; dellog = new double[n]; int npos = 0; for (int i = 0; i < n; i++) { if (t.getElement(i) > 0. && y.getElement(i) > 0.) { tlog[npos] = Math.log(t.getElement(i)); ylog[npos] = Math.log(y.getElement(i)); dellog[npos] = 1.; npos++; } } DatanVector vtlog = new DatanVector(npos); DatanVector vylog = new DatanVector(npos); DatanVector vdellog = new DatanVector(npos); for (int j = 0; j < npos; j++) { vtlog.setElement(j, tlog[j]); vylog.setElement(j, ylog[j]); vdellog.setElement(j, dellog[j]); } LsqPol lp = new LsqPol(vtlog, vylog, vdellog, 2); x = lp.getResult(); x.setElement(0, Math.exp(x.getElement(0))); df.writeLine(" x = " + x.toString()); // perform fit int[] list = {1, 1}; powerlaw = new Powerlaw(); LsqNon ln = new LsqNon(t, y, dy, x, list, powerlaw); x = ln.getResult(); x1 = x.getElement(0); x2 = x.getElement(1); cov = ln.getCovarianceMatrix(); delx1 = Math.sqrt(cov.getElement(0, 0)); delx2 = Math.sqrt(cov.getElement(1, 1)); rho = cov.getElement(1, 0) / (delx1 * delx2); m = ln.getChiSquare(); p = 1. - StatFunct.cumulativeChiSquared(m, n - 1); // curve of fitted exponential xpl = new double[1001]; ypl = new double[1001]; dpl = (double) (n - 1) * deltat / 1000.; for (int i = 0; i < 1001; i++) { xpl[i] = t0 + (double) i * dpl; ypl[i] = x1 * Math.pow(xpl[i], x2); } // prepare data points for plotting datx = new double[n]; daty = new double[n]; datsx = new double[n]; datsy = new double[n]; datrho = new double[n]; for (int i = 0; i < n; i++) { datx[i] = t.getElement(i); daty[i] = y.getElement(i); datsx[i] = 0.; datsy[i] = dy.getElement(i); datrho[i] = 0.; } // display data and fitted curve caption = "x_1#=" + String.format(Locale.US, "%5.2f", x1) + ", x_2#=" + String.format(Locale.US, "%5.2f", x2) + ", &D@x_1#=" + String.format(Locale.US, "%5.2f", delx1) + ", &D@x_2#=" + String.format(Locale.US, "%5.2f", delx2) + ", &r@=" + String.format(Locale.US, "%5.2f", rho) + ", M=" + String.format(Locale.US, "%5.2f", m) + ", P=" + String.format(Locale.US, "%6.4f", p); new GraphicsWithDataPointsAndPolyline( getClass().getName(), "", xpl, ypl, 1, .3, datx, daty, datsx, datsy, datrho, "t", "y", caption); }