I'm trying to conduct the residual analysis for simple linear regression. I need to prove that the residuals follow an approximate Normal Distribution. The csv file I'm using has values for Percentage of marks in Grade 10 and the Salary the student makes. Once I run the below code, my plot looks like this: The plot in the book looks like this: I was expecting my plot to show up like the book as the data is the same. I have double-checked to make sure I'm not missing any data etc. I have split the data set into training and test as per the book as well. Data is as follows: Percentage Salary 62 270000 76.33 200000 72 240000 60 250000 61 180000 55 300000 70 260000 68 235000 82.8 425000 59 240000 58 250000 60 180000 66 428000 83 450000 68 300000 37.33 240000 79 252000 68.4 280000 70 231000 59 224000 63 120000 50 260000 69 300000 52 120000 49 120000 64.6 250000 50 180000 74 218000 58 360000 67 150000 75 250000 60 200000 55 300000 78 330000 50.08 265000 56 340000 68 177600 52 236000 54 265000 52 200000 76 393000 64.8 360000 74.4 300000 74.5 250000 73.5 360000 57.58 180000 68 180000 69 270000 66 240000 60.8 300000 The code is below: # Importing all required libraries for building the regression model import pandas as pd import numpy as np import statsmodels.api as sm from sklearn.model_selection import train_test_split # Load the dataset into dataframe mba_salary_df = pd.read_csv( 'MBA Salary.csv' ) # Add constant term of 1 to the dataset X = sm.add_constant( mba_salary_df[‘Percentage in Grade 10’] ) # Split dataset into train and test set into 80:20 respectively train_X, test_X, train_y, test_y = train_test_split( X, Y, train_size = 0.8,random_state = 100 ) # Fit the regression model mba_salary_lm = sm.OLS( train_y, train_X ).fit() mba_salary_resid = mba_salary_lm.resid probplot = sm.ProbPlot(mba_salary_resid) plt.figure( figsize = (8, 6) ) probplot.ppplot(line='45') plt.title("Normal P-P Plot of Regression Standardized Residuals") plt.show() Continue reading...
