diff --git a/Kuengjoe_S04/HM2_Serie03.pdf b/Kuengjoe_S04/HM2_Serie03.pdf
new file mode 100644
index 0000000..ce59537
Binary files /dev/null and b/Kuengjoe_S04/HM2_Serie03.pdf differ
diff --git a/Kuengjoe_S04/Kuengjoe_S04_Aufg2.py b/Kuengjoe_S04/Kuengjoe_S04_Aufg2.py
new file mode 100644
index 0000000..fcc2577
--- /dev/null
+++ b/Kuengjoe_S04/Kuengjoe_S04_Aufg2.py
@@ -0,0 +1,22 @@
+import numpy as np
+
+def lagrange_int(x, y, x_int):
+    s = 0
+
+    for i in range(len(x)):
+        L = 1
+        for j in range(len(x)):
+            if i != j:
+                L = L * (x_int - x[j]) / (x[i] - x[j])
+        s = s + y[i] * L
+
+    return s
+
+
+x = [0, 2500, 5000, 10000]
+y = [1013, 747, 540, 226]
+x_int = 3750
+
+y_int = lagrange_int(x, y, x_int)
+
+print("y_int =", y_int)
\ No newline at end of file
diff --git a/Kuengjoe_S04/Kuengjoe_S04_Aufg3.py b/Kuengjoe_S04/Kuengjoe_S04_Aufg3.py
new file mode 100644
index 0000000..0d6efd3
--- /dev/null
+++ b/Kuengjoe_S04/Kuengjoe_S04_Aufg3.py
@@ -0,0 +1,82 @@
+import numpy as np
+import matplotlib.pyplot as plt
+
+x = np.array([1981, 1984, 1989, 1993, 1997, 2000, 2001, 2003, 2004, 2010], dtype=float)
+y = np.array([0.5, 8.2, 15, 22.9, 36.6, 51, 56.3, 61.8, 65, 76.7], dtype=float)
+
+x_plot = np.arange(1975, 2020.1, 0.1)
+
+p1 = np.polyfit(x, y, len(x) - 1)
+y_plot1 = np.polyval(p1, x_plot)
+
+plt.figure()
+plt.plot(x_plot, y_plot1)
+plt.plot(x, y, "o")
+plt.xlim(1975, 2020)
+plt.ylim(-100, 250)
+plt.title("Aufgabe 3a")
+plt.grid()
+
+# Das Polynom geht praktisch durch alle Datenpunkte.
+
+
+
+x_mean = np.mean(x)
+x_shift = x - x_mean
+x_plot_shift = x_plot - x_mean
+
+p2 = np.polyfit(x_shift, y, len(x_shift) - 1)
+y_plot2 = np.polyval(p2, x_plot_shift)
+
+plt.figure()
+plt.plot(x_plot, y_plot2)
+plt.plot(x, y, "o")
+plt.xlim(1975, 2020)
+plt.ylim(-100, 250)
+plt.title("Aufgabe 3b")
+plt.grid()
+
+# Mit verschobenen x-Werten ist die numerische Stabilität besser.
+
+
+
+y_2020_a = np.polyval(p1, 2020)
+y_2020_b = np.polyval(p2, 2020 - x_mean)
+
+print("Schätzwert 2020 aus a):", y_2020_a)
+print("Schätzwert 2020 aus b):", y_2020_b)
+
+# Der Schätzwert für 2020 ist unsicher, weil er außerhalb des Datenintervalls liegt.
+
+
+
+def lagrange_int(x, y, x_int):
+    y_int = np.zeros(len(x_int))
+
+    for k in range(len(x_int)):
+        s = 0
+        for i in range(len(x)):
+            L = 1
+            for j in range(len(x)):
+                if i != j:
+                    L = L * (x_int[k] - x[j]) / (x[i] - x[j])
+            s = s + y[i] * L
+        y_int[k] = s
+
+    return y_int
+
+y_lagrange = lagrange_int(x, y, x_plot)
+
+plt.figure()
+plt.plot(x_plot, y_lagrange, label="Lagrange")
+plt.plot(x_plot, y_plot2, label="polyfit b)")
+plt.plot(x, y, "o")
+plt.xlim(1975, 2020)
+plt.ylim(-100, 250)
+plt.title("Aufgabe 3d")
+plt.grid()
+plt.legend()
+
+# Beide Kurven sind fast gleich. Kleine Unterschiede kommen von numerischen Fehlern.
+
+plt.show()
\ No newline at end of file