Python Machine Learning: Linear Regression (II)

 In the previous tutorial, you have learned how to build a linear regression using matrix multiplication (please go to Python Machine Learning: Linear Regression (I)). Now, in this tutorial, a Machine Learning Python library called scikit-learn will be used for this purpose. 

Once we have imported the data from the text file, let's set our x- and y-values. 


#Importing libraries

import numpy as np


#Importing text file

data = np.loadtxt('points.txt', skiprows=(2), dtype=float)

print(data)


#Setting x values

x = data[:,0]

print(x)


#Setting y values

y = data[:,1]

print(y)




From the figure above (an extract of the whole data), we can notice that x and y are 1D array. If we want to work with the scikit-learn Machine Learning Python library, it is necessary to convert our 1D arrays into 2D. For this, the function reshape(-1,1).


#Reshaping the array into a vector-column

x2 = data[:,0].reshape(-1,1)

print(x2)


#Reshaping the array into a vector-column

y2 = data[:,1].reshape(-1,1)

print(y2)




Now, we are able to build our linear regression model using the LinearRegression module from the scikit-learn library. Do not forget to import the library.


#Importing library

from sklearn.linear_model import LinearRegression


#Building the linear regression model

linear_regression = LinearRegression()

linear_model = linear_regression.fit(x2,y2)


As explained in the previous tutorial, the linear relationship can be as y = c0 + c1*x, where c0 is the intercept with the y-axis, and c1 is the slope of the line. These two coefficients can be found easier and faster thanks to the function LinearRegression().fit( ).  In order to get both coefficients, the functions intercept_ and coef_ are needed.


#Getting the intercept with y-axis

intercept_yaxis = linear_model.intercept_

print(intercept_yaxis)


#Getting the coefficient

slope = linear_model.coef_

print(slope)




In contrast to the matrix multiplication approach where the coefficient matrix is an array of two elements, both elements are now got in two different arrays of one element each. If comparing both approaches, both intercept and slope should be exactly the same. The coefficient matrix from the previous tutorial was the following:




As seen from both pictures, we notice that both coefficients (intercept and slope) are exactly the same. This means we did a great job making the linear regression! Finally, let's establish the linear relationship and plot it.


#Importing library

import matplotlib.pyplot as plt


#Establishing the linear relationship

y_lineal2 = slope*x2 + intercept_yaxis

print(y_lineal2)


#Plotting


#Initially given x- and y-points

plt.scatter(x,y)

#Linear regression points

plt.plot(x2, y_lineal2, color='red')

#Naming the graph, x- and y-axis

plt.title('scikit-learn library')

plt.xlabel('x')

plt.ylabel('y')

plt.show()




The plot we got in the previous tutorial was the following:



As seen from both graphics, we can say they are exactly the same! The final Python code will look like the following:



Congratulations! You just made your first Machine Learning regression. In the next tutorial, polynomial regression will be explained.

Comments

Post a Comment

Popular posts from this blog

Python: Tracking any phone number

Image
 Did you know Python is capable of getting the location of any person just by knowing his/her phone number? In this exercise, we will see how this is possible. Once again, this tutorial is only for educational and entertainment purposes. The libraries geocoder , and carrier libraries from the  phonenumbers  library should be imported. These libraries will help with the purpose of this exercise. Let's start! import phonenumbers from phonenumbers import geocoder from phonenumbers import carrier #Introducing the phone number to be tracked number = '+51 XXXXXXXXX' phone_number = phonenumbers.parse(number) carrier = carrier.name_for_number(phone_number, 'es') region = geocoder.description_for_number(phone_number,'es') As seen from the pictures, Python can output the name of the region in any language. If you want, for example, in English, you write 'en' instead of 'es', or 'de' for German, or 'ru' for Russian. The output of the va...
Read more

Python: Pandas DataFrame data manipulation

Image
 There are some ways to manipulate data in Python. One method is using the library p andas for DataFrame and another method is using another library named numpy for array. In this section, you will learn data manipulation using DataFrame. The library  p andas allows visualizing a vast amount of data in a very nice way. You are able to work with data that looks like data in any Excel sheet. For this exercise, we will use the text file named 'vaccinations.txt' from the previous exercise. Let's get started! Just to remind you, in the previous tutorial (please go to Python: Downloading data from the web ) , we downloaded a file from the web, reading line by line, and finally, created a text file named 'vaccinations.txt', containing a vast amount of  data about covid vaccination in the world. This file will be used in this exercise. A very short part of it looks like this: As seen from the figure above, the columns are separated by the  'comma' (',') deli...
Read more

Python Machine Learning: Linear Regression (I)

Image
Have you ever felt your phone hears your conversation? For example, you and your friend are talking about new shoes, and then when you pick your phone up, you get a bunch of ads about it? Or, when you watch a movie or series on Netflix, the next time you get recommendations of your taste?  Well, this all is possible thanks to Machine Learning! If you have not heard about it, let me explain it to you. Machine Learning is a very popular topic nowadays. It is a method to analyze data in analytical performance. Machine Learning helps humans with very complicated topics, such as forecasting bitcoin price. In machine learning, the AI model learns from data, analyzes it, and, then, establishes patterns to make future decisions. In this and in the following tutorials, you will learn the basics of Machine Learning using Python. In this tutorial, linear regression will be explained. Before we start coding, what is linear regression? Well, linear regression is an algorithm where the predicted...
Read more