Polynomial regression

In [1]:

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt

In [2]:

dataset = pd.read_csv('Position_Salaries.csv')
dataset.head(2)

Out[2]:

	Position	Level	Salary
0	Business Analyst	1	45000
1	Junior Consultant	2	50000

In [3]:

x = dataset.iloc[:,[1]].values
y = dataset.iloc[:,[2]].values

In [4]:

#from sklearn.cross_validation import train_test_split
#x_train,x_test,y_train,y_test = train_test_split(x,y)

In [5]:

#normal linear regression
from sklearn.linear_model import LinearRegression
reg = LinearRegression()
reg.fit(x,y)
y_pred = reg.predict(x)
plt.scatter(x,y,color = 'r')
plt.scatter(x,reg.predict(x),color = 'g')
plt.plot(x,y_pred)
plt.plot(x,y)

Out[5]:

[<matplotlib.lines.Line2D at 0x1bb22c81b00>]

In [6]:

#polynomial linear regression
from sklearn.preprocessing import PolynomialFeatures  #to make poly feature from 1 column to degree+1 column
from sklearn.linear_model import LinearRegression  
reg1 = LinearRegression()
poly_reg = PolynomialFeatures(degree = 3)
x_poly = poly_reg.fit_transform(x)
reg1.fit(x_poly,y)

plt.scatter(x,y,color = 'r')
plt.scatter(x,reg1.predict(x_poly),color = 'g')
plt.plot(x,reg1.predict(x_poly),color = 'b')

Out[6]:

[<matplotlib.lines.Line2D at 0x1bb2301dfd0>]

In [7]:

print(reg.predict(6.5)) #linear regression
print(reg1.predict(poly_reg.fit_transform(6.5))) #polynomial linear regression

[[330378.78787879]]
[[133259.46969697]]

In [ ]:

In [ ]: