Added regression work using sklearn lib

Author: garrysingh2930

linear-algebra-ipynb/Linear_systems_with_2_variables.ipynb renamed to linear-algebra-ml-ipynb/Linear_systems_with_2_variables.ipynb

@@ -0,0 +1,65 @@
+# -*- coding: utf-8 -*-
+"""
+Regression ML SkLearn Python script
+Author: Garry Singh
+
+This script covers regression computations in python using Scikit Learn library
+
+"""
+
+# Importing necessary libraries
+import numpy as np
+from sklearn.linear_model import SGDRegressor # lib for computing gradient descent
+from sklearn.preprocessing import StandardScaler # lib for computing z score normalized features (feature scaling)
+import matplotlib.pyplot as plt
+
+# let's prepare sample input data
+
+# Create a NumPy array with feature names as strings
+x_features = np.array(["Square Feet", "Bedrooms", "Bathrooms", "Neighborhood", "Garage"])
+
+# Sample input data as a 2D NumPy array
+x_train = np.array([
+    [1500, 3, 2, 1, 1],
+    [1800, 4, 2.5, 2, 0],
+    [1200, 2, 1, 0, 0],
+    [2100, 4, 3, 1, 1],
+    [1600, 3, 2.5, 2, 1],
+    [1400, 2, 1.5, 1, 0],
+    [2500, 4, 3.5, 2, 1]
+])
+
+# Corresponding prices as a 1D NumPy array
+y_train = np.array([250000, 320000, 180000, 400000, 300000, 220000, 450000])
+
+# let's normalize the given input training data
+scalar = StandardScaler()
+x_norm = scalar.fit_transform(x_train)
+
+#print(f"Normalized housing feature data is: \n {x_norm}")
+
+# let's try to fit the normalized feature data with given price training values
+sgdr = SGDRegressor(max_iter=100000)
+sgdr.fit(x_norm, y_train)
+
+n_iter_ran = sgdr.n_iter_
+n_weight_updates = sgdr.t_
+b_norm = sgdr.intercept_ 
+w_norm = sgdr.coef_
+
+#print(f"Number of iterations used: {n_iter_ran}")
+#print(f"Number of updates completed to weight: {n_weight_updates}")
+
+#print(f"Normalized bias is: {b_norm} ")
+#print(f"Normalized weight is: {w_norm} ")
+
+# now let's try to predict price for given denormalized data
+
+y_pred_sgd = sgdr.predict(x_norm)
+y_pred_dot = np.dot(x_norm, w_norm) + b_norm
+
+for i in range(x_train.shape[1]):
+    plt.scatter(x_train[:,i], y_train,marker='o', color='g')
+    plt.ylabel("House price in $")
+    plt.xlabel(x_features[i])
+    plt.show()