From eabe17cc1d90919c8484592ba60f19758508efad Mon Sep 17 00:00:00 2001
From: Soham Haldar <115897153+Mahos-H@users.noreply.github.com>
Date: Sat, 19 Oct 2024 11:16:43 +0530
Subject: [PATCH] Add files via upload

Perceptron.py shows a simple Perceptron model capable of backpropagation and how exactly the weights are updated and visualization of the same, considering 2 input nodes and 1 output nodes. Activation function is sigmoid and loss function is mean squared error.
---
 Python/Perceptron.py | 89 ++++++++++++++++++++++++++++++++++++++++++++
 1 file changed, 89 insertions(+)
 create mode 100644 Python/Perceptron.py

diff --git a/Python/Perceptron.py b/Python/Perceptron.py
new file mode 100644
index 000000000..6c0d6df4a
--- /dev/null
+++ b/Python/Perceptron.py
@@ -0,0 +1,89 @@
+import numpy as np
+import matplotlib.pyplot as plt
+from matplotlib.animation import FuncAnimation
+
+class Perceptron:
+    def __init__(self, learning_rate=0.1, max_iters=100):
+        # Initialize parameters
+        self.learning_rate = learning_rate
+        self.max_iters = max_iters
+        self.weights = None
+        self.bias = None
+        self.history = []
+
+    def sigmoid(self, x):
+        # Sigmoid activation function
+        return 1 / (1 + np.exp(-x))
+
+    def fit(self, X, y):
+        # Initialize weights and bias
+        n_samples, n_features = X.shape
+        self.weights = np.random.randn(n_features)
+        self.bias = 1
+
+        for _ in range(self.max_iters):
+            correct_classifications = True  # Track if all predictions are correct
+
+            for i, x_i in enumerate(X):
+                # Compute linear combination and apply sigmoid
+                linear_output = np.dot(x_i, self.weights) + self.bias
+                y_pred = self.sigmoid(linear_output)
+
+                # Classify based on sigmoid output
+                predicted_label = 1 if y_pred >= 0.5 else 0
+
+                # If misclassified, update weights and bias
+                if predicted_label != y[i]:
+                    correct_classifications = False
+                    error = y[i] - y_pred
+                    self.weights += self.learning_rate * error * x_i
+                    self.bias += self.learning_rate * error
+
+                # Store history for visualization
+                self.history.append((self.weights.copy(), self.bias))
+
+            # Stop if all samples are classified correctly
+            if correct_classifications:
+                break
+
+    def get_history(self):
+        # Return history of updates
+        return self.history
+
+def animate(i):
+    # Get weights and bias for the current iteration
+    weights, bias = history[i]
+    ax.clear()
+    
+    # Plot data points
+    for j in range(len(y)):
+        if y[j] == 1:
+            ax.scatter(X[j, 0], X[j, 1], marker='o', color='blue', s=100)
+        else:
+            ax.scatter(X[j, 0], X[j, 1], marker='x', color='red', s=100)
+
+    # Plot decision boundary
+    x_values = np.linspace(min(X[:, 0])-1, max(X[:, 0])+1, 100)
+    y_values = -(weights[0] * x_values + bias) / weights[1]
+    ax.plot(x_values, y_values, color='green')
+
+    ax.set_xlim(min(X[:, 0])-2, max(X[:, 0])+2)
+    ax.set_ylim(min(X[:, 1])-2, max(X[:, 1])+2)
+    ax.set_title(f'Iteration {i+1}')
+
+# Training data
+X = np.array([[2, 1], [-2, -1], [1, -2], [-1, -2], [2, -1], [-1, 1], [3, 4]])
+y = np.array([1, 0, 1, 0, 1, 0, 0])
+
+# Train perceptron model
+perceptron = Perceptron(learning_rate=0.1, max_iters=100)
+perceptron.fit(X, y)
+
+# Get history of weight updates for animation
+history = perceptron.get_history()
+
+# Set up plot for animation
+fig, ax = plt.subplots()
+plt.grid(True)
+ani = FuncAnimation(fig, animate, frames=len(history), interval=500, repeat=False)
+plt.show()