Linear Regression's coefficient and intercept kept changing

Question

bach_sal = bach['Salary']
masters_sal = masters['Salary']
phd_sal = phd['Salary']
deg_category_sal = [assoc_sal, bach_sal, masters_sal, phd_sal]

x = np.array(assoc['Person'])
y = np.array(assoc_sal)
from sklearn.model_selection import train_test_split
x_train, x_test, y_train, y_test = train_test_split(x,y,test_size=0.7)
from sklearn.linear_model import LinearRegression
lr = LinearRegression()
lr.fit(x_train.reshape(-1,1),y_train)
print (lr.coef_)
print (lr.intercept_)

I received the coefficient and y intercept but everytime I run the code, they change to a different number. Please help. Thank you.

Answer 1

This is because you did not set the random_state in the train_test split function. This shuffles the cut for test data on each run and as a result, the coefficient changes slightly. Set the random_state to an integer to get the same answer over multiple runs.

https://scikit-learn.org/stable/modules/generated/sklearn.model_selection.train_test_split.html

Answer 2

This will keep changing until you specify a fixed value of random_state to your train test line number.

x_train, x_test, y_train, y_test = train_test_split(x,y,test_size=0.3, random_state=42)

After making this change you might get exact same values. In general, your train and test data will keep on changing until you specify set rows for training and test.