Linear Model in Python

Basic Work Flow:

1. check distribution for each variable for assumptions
2. check for outliers and any transformation need to perform (log?)
3. fit a linear regression model
4. Analyze model fitting (R square)
5. predict (un-transform the result if transformation is performed)

# useful packages
import pandas as pd
import json
import statistics
import urllib
import numpy as np
import scipy
import matplotlib.pyplot as plt

# assume features are log-trandformed
X = data["feature1", "feature2"]
X = sm.add_constant(X)
y = data["outcome"]

model = sm.OLS(y, X)
result = model.fit()

print("R Squared is:", round(result.rsquared,3))

Xpred = np.log(range(8, 11))
Xpred = sm.add_constant(Xpred)
print("Predicted results for 10 is :",
  round(np.exp(result.predict(Xpred))[2], 3))

Advertisement