test_code_from_uny.py

# IU - International University of Applied Science
# Machine Learning - Supervised Learning
# Course Code: DLBDSMLSL01

# Generalized Linear Model (GLM)

# %% load packages
import statsmodels.api as sm
import pandas as pd

# %% load data
dataset = pd.read_csv('datasets/Salary_Data.csv')
x = dataset.drop(columns=['Salary'])
y = dataset.iloc[:, 1].values

# %% specify exogeneous and endogeneous variables
exog, endog = sm.add_constant(x), y

# %% specify the model
mod = sm.GLM(
    endog,
    exog,
    family=sm.families.Poisson(
        link=sm.families.links.log()
    )
)

# %% fit the model
res = mod.fit()

# %% print model summary
print(res.summary())



import matplotlib.pyplot as plt

# Generate predictions
predicted = res.predict(exog)

# Plot actual points and predicted line
plt.scatter(x, y, color='blue', label='Actual')
plt.plot(x, predicted, color='red', label='Predicted')
plt.xlabel('Years of Experience')
plt.ylabel('Salary')
plt.title('GLM: Actual vs Predicted')
plt.legend()
plt.savefig('glm_plot_iu_example.png')  # Save the plot
plt.show()