Linear Regression is a method or approach for Supervised Learning.Supervised Learning takes the historical or past data and then train the model and predict the things according to the past results.Linear Regression comes from the word ‘Linear’ and ‘Regression’.Regression concept deals with predicting the future using the past data.Linear means the able to represented by a straight line on graph.Linear Regression has two things one independent variable and other dependent variable and Linear Regression is a relationship between the two.
In this article, we are going to learn about how we can implement Linear Regression with the help of Turicreate. Turicreate is Library in Python which helps the beginners to learn and implement Machine Learning Algorithm easily as well as efficiently.
Step 1: Importing the Turicreate Library
Python3
import turicreate as tc |
Step 2: Reading the datasets.
Python3
"""The good thing about Turicreate is that we don't haveimport any other library for data loading.Turicreate itself can load data with the help of it's SFrame Data Structure"""data_sets = tc.SFrame("data.csv") |
Link for the data is=https://www.kaggle.com/mirichoi0218/insurance
Step 3: Exploring the data
Python3
# It will display the first few Lines of the datadata.head() |
Output:
datasets first few lines
Step 4: Make a Linear Regression model .
Python3
# We will have a target variable that stores the thing# to the predicted and feature is the list of elements# which we will take for making the model.model = tc.linear_regression.create( data, target ="charges", features =['region']) |
Step 5: Now evaluate the model
Python3
# this will tell us about the max error and # the rmse (root mean squared error)model.evaluate(data) |
Output:
Max error and Rmse
Step 6: Now predicting the charges according to the B.M.I of person
Python3
# this variable will be containing the data# of person having the bmi 27.9bmi_person = data[data['bmi']== 27.9]# it will predict the charges for the person with bmi 27.9model.predict(bmi_person) |
Output:
The Prediction of the charges
