Page 334 - AI Ver 3.0 Class 11
P. 334

Advantage of Linear Regression

              Some advantages of linear regression are as follows:
              ●  Linear regression is a simple technique and easy to implement.
              ●  Efficient to train the machine on this model.


              Disadvantages of Linear Regression
              Some disadvantages of linear regression are as follows:
              ●  Regression analysis is sensitive to outliers as these can have a great impact on the analysis.
              ●   It is quite prone to overfitting. (Overfitting means that the training of the model on data is just too good and the
                  test sample size is quite small).

                             Reboot


                    1.   State the two types of Regression.
                    2.  How many variables are used in linear regression?

                    3.  State the equation of the line of best fit.

                    4.  Why is it called the line of best fit?
                    5.  State two applications of regression.





              For Advanced Learners
                Program 1: To demonstrate the use of simple linear regression in Python

              import numpy as np
              import matplotlib.pyplot as plt



              # Updated sample data with more values
              data_x = np.array([2, 4, 8, 6, 8, 10, 14, 12, 16, 20])
              data_y = np.array([3, 5, 7, 8, 9, 6, 7, 5, 9, 8])


              # Calculate mean and standard deviation

              mean_x = np.mean(data_x)
              mean_y = np.mean(data_y)
              std_x = np.std(data_x)
              std_y = np.std(data_y)


              # Calculate covariance and slope
              cov = np.sum((data_x - mean_x) * (data_y - mean_y)) / (len(data_x) - 1)
              slope = cov / (std_x**2)



              # Calculate y-intercept (b)
              intercept = mean_y - slope * mean_x
                    332     Touchpad Artificial Intelligence (Ver. 3.0)-XI
   329   330   331   332   333   334   335   336   337   338   339