#840 adds polymonial regression program in python

machine_learning/polymonial_regression.py

@@ -0,0 +1,38 @@
+import numpy as np
+import matplotlib.pyplot as plt
+import pandas as pd
+
+# Importing the dataset
+dataset = pd.read_csv('https://s3.us-west-2.amazonaws.com/public.gamelab.fun/dataset/position_salaries.csv')
+X = dataset.iloc[:, 1:2].values
+y = dataset.iloc[:, 2].values
+
+
+# Splitting the dataset into the Training set and Test set
+from sklearn.model_selection import train_test_split 
+X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=0)
+
+
+# Fitting Polynomial Regression to the dataset
+from sklearn.preprocessing import PolynomialFeatures
+from sklearn.linear_model import LinearRegression
+poly_reg = PolynomialFeatures(degree=4)
+X_poly = poly_reg.fit_transform(X)
+pol_reg = LinearRegression()
+pol_reg.fit(X_poly, y)
+
+
+# Visualizing the Polymonial Regression results
+def viz_polymonial():
+    plt.scatter(X, y, color='red')
+    plt.plot(X, pol_reg.predict(poly_reg.fit_transform(X)), color='blue')
+    plt.title('Truth or Bluff (Linear Regression)')
+    plt.xlabel('Position level')
+    plt.ylabel('Salary')
+    plt.show()
+    return
+viz_polymonial()
+
+# Predicting a new result with Polymonial Regression
+pol_reg.predict(poly_reg.fit_transform([[5.5]]))
+#output should be 132148.43750003