[Python] Why are my VIF scores so low, but in my OLS Regression Results, I get a warning: the...

mport pandas as pd import numpy as np import matplotlib.pyplot as plt import statsmodels.api as sm import seaborn as sns from statsmodels.stats.outliers_influence import variance_inflation_factor from sklearn.model_selection import train_test_split

def remove_dot(data): dot_not_accepted = ['brthord', 'meduc', 'feduc'] for column in dot_not_accepted: data[column] = data[column].replace('.', np.nan) data = data.dropna() return data

data_file_path = 'data.xls'

Load your data
data = pd.read_excel(data_file_path)

Process the data using the remove_dot function
data = remove_dot(data)

Create the regression model
data['Total_Education'] = data['feduc'] + data['meduc'] X = data[['hours', 'IQ', 'KWW', 'educ', 'exper', 'tenure', 'age', 'sibs','brthord', 'feduc', 'meduc']] y = data['wage'] X = sm.add_constant(X) # Adds a constant to the model

Split the data into training and test sets
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=0)

Fit the model with training data
model = sm.OLS(y_train, X_train).fit()

Calculate VIF for each variable
vif_data = pd.DataFrame() vif_data["Variables"] = X_train.columns vif_data["VIF"] = [variance_inflation_factor(X_train.values, i) for i in range(X_train.shape[1])] print(vif_data)

Print the summary of the model
print(model.summary())

Variables VIF 0 const 249.241998 1 hours 1.043743 2 IQ 1.564278 3 KWW 1.732137 4 educ 2.061102 5 exper 1.942750 6 tenure 1.183890 7 age 1.800178 8 sibs 1.515675 9 brthord 1.465356 10 feduc 1.622698 11 meduc 1.620933 Notes: [1] Standard Errors assume that the covariance matrix of the errors is correctly specified. [2] The condition number is large, 1.98e+03. This might indicate that there are strong multicollinearity or other numerical problems.
​

Is it an error in my code? Or is this normal behaviour.

I tried to find a fault in my code, because I find my case very suspicious. I have to find the best model for the data. I thought it was a nice idea, that I would combine or delete variables based on VIF score. But now is everything according to VIF not multicollinear.

Continue reading...