Add perceptron basic model from scratch

Perceptron/Perceptron.py

@@ -0,0 +1,28 @@
+import numpy as np 
+
+class Perceptron:
+	def __init__(self, learning_rate = 0.001, n_iter = 10000):
+		self.learning_rate = learning_rate
+		self.n_iter = n_iter
+		self.activation_function = unit_step_function
+		self.weights = None
+		self.bias = None
+
+	def unit_step_function(self, X):
+		return np.where(X >= 0, 1, 0)
+
+	def fit(self, X, Y):
+		n_samples, n_features = X.shape
+		self.weights = np.zeros(n_features)
+		self.bias = 0
+		Y = np.array([1 if x > 0 else 0 for x in Y])
+		for i in range(self.n_iter):
+			for idx, x in enumerate(X):
+				solve = np.dot(x, self.weights) + self.bias
+				y_pred = self.activation_function(solve)
+				self.weights += self.learning_rate * (Y[idx] - y_pred) * x
+				self.bias += self.learning_rate * (Y[idx] - y_pred)
+
+	def predict(self, X):
+		solve = np.dot(X, self.weights) + self.bias
+		return self.activation_function(solve)