untracked m-files

programming/sinusAmplituden.m

@@ -0,0 +1,8 @@
+x = (0:0.01:2*pi);
+frequenz = 1.0;
+amplituden = [0.25 0.5 0.75 1.0 1.25];
+hold on
+for i = 1:length(amplituden)
+    y = sin(frequenz * x) * amplituden(i);
+    plot(x, y)
+end

programming/sinusFrequenzen.m

@@ -0,0 +1,6 @@
+frequenzen = [1.0 2.0 4.0];
+
+for j = 1:length(frequenzen)
+    frequenz = frequenzen(j);
+    sinusAmplituden
+end

statistics/do_regression.m

@@ -0,0 +1,6 @@
+function parameter = do_regression(x, y, start_parameter)
+
+options = optimoptions('fminunc','GradObj','on');
+fun = @exp_funct;
+myfunc = @ (p)(lserr(p, x, y, fun));
+parameter = fminunc(myfunc, start_parameter, options);

statistics/exp_funct.m

@@ -0,0 +1,3 @@
+function y = exp_funct(A, tau, t)
+
+y = A .* (1 - exp(-1 * (t./tau)));

statistics/lin_funct.m

@@ -0,0 +1,3 @@
+function y = lin_funct(a, b, x)
+
+y = a.*x + b;

statistics/lserr.m

@@ -0,0 +1,7 @@
+function [err, grad] = lserr(p, x, y, fun)
+
+err = sum(y - fun(p(1), p(2), x)).^2;
+
+if nargout > 1
+    grad = lserr_gradient(p, x, y);
+end

statistics/lserr_gradient.m

@@ -0,0 +1,10 @@
+function grad = lserr_gradient(param, x, y)
+    h = 1e-6;
+    
+    grad = 0*param;
+    for i = 1:length(param)
+        paramh = param;
+        paramh(i) = param(i) + h;
+        grad(i) = (lserr(paramh,x,y) - lserr(param,x,y))/h;
+    end
+