A Guide to Understanding Interaction Terms

Introduction

Interaction terms are incorporated in regression modelling to capture the effect of two or more independent variables in the dependent variable. At times, it is not just the simple relationship between the control variables and the target variable that is under investigation, interaction terms can be quite helpful at these moments. These are also useful whenever the relationship between one independent variable and the dependent variable is conditional on the level of another independent variable.

This, of course, implies that the effect of one predictor on the response variable depends on the level of another predictor. In this blog, we examine the idea of interaction terms through a simulated scenario: predicting time and again the amount of time users would spend on an e-commerce channel using their past behavior.

Learning Objectives

• Understand how interaction terms enhance the predictive power of regression models.
• Learn to create and incorporate interaction terms in a regression analysis.
• Analyze the impact of interaction terms on model accuracy through a practical example.
• Visualize and interpret the effects of interaction terms on predicted outcomes.
• Gain insights into when and why to apply interaction terms in real-world scenarios.

This article was published as a part of the Data Science Blogathon.

Understanding the Basics of Interaction Terms

In real life, we do not find that a variable works in isolation of the others and hence the real-life models are much more complex than those that we study in classes. For example, the effect of the end user navigation actions such as adding items to a cart on the time spent on an e-commerce platform differs when the user adds the item to a cart and buys them. Thus, adding interaction terms as variables to a regression model allows to acknowledge these intersections and, therefore, enhance the model’s fitness for purpose in terms of explaining the patterns underlying the observed data and/or predicting future values of the dependent variable.

Mathematical Representation

Let’s consider a linear regression model with two independent variables, X1​ and X2:

Y = β0​ + β1​X1​ + β2​X2​ + ϵ,

where Y is the dependent variable, β0​ is the intercept, β1​ and β2​ are the coefficients for the independent variables X1​ and X2, respectively, and ϵ is the error term.

Adding an Interaction Term

To include an interaction term between X1​ and X2​, we introduce a new variable X1⋅X2 ​:

Y = β0 + β1X1 + β2X2 + β3(X1⋅X2) + ϵ,

where β3 represents the interaction effect between X1​ and X2​. The term X1⋅X2 is the product of the two independent variables.

How Interaction Terms Influence Regression Coefficients?

• β0​: The intercept, representing the expected value of Y when all independent variables are zero.
• β1​: The effect of X1​ on Y when X2​ is zero.
• β2​: The effect of X2​ on Y when X1​ is zero.
• β3​: The change in the effect of X1​ on Y for a one-unit change in X2​, or equivalently, the change in the effect of X2​ on Y for a one-unit change in X1.​

Example: User Activity and Time Spent

First, let’s create a simulated dataset to represent user behavior on an online store. The data consists of:

• added_in_cart: Indicates if a user has added products to their cart (1 for adding and 0 for not adding).
• purchased: Whether or not the user completed a purchase (1 for completion or 0 for non-completion).
• time_spent: The amount of time a user spent on an e-commerce platform. Our goal is to predict the duration of a user’s visit on an online store by analysing if they add products to their cart and complete a transaction.

# import libraries
import pandas as pd
import numpy as np

# Generate synthetic data
def generate_synthetic_data(n_samples=2000):

    np.random.seed(42)
    added_in_cart = np.random.randint(0, 2, n_samples)
    purchased = np.random.randint(0, 2, n_samples)
    time_spent = 3 + 2*purchased + 2.5*added_in_cart + 4*purchased*added_in_cart + np.random.normal(0, 1, n_samples)
    return pd.DataFrame({'purchased': purchased, 'added_in_cart': added_in_cart, 'time_spent': time_spent})

df = generate_synthetic_data()
df.head()

Output:

Simulated Scenario: User Behavior on an E-Commerce Platform

As our next step we will first build an ordinary least square regression model with consideration to these actions of the market but without coverage to their interaction effects. Our hypotheses are as follows: (Hypothesis 1) There is an effect of the time spent on the website where each action is taken separately. Now we will then construct a second model that includes the interaction term that exists between adding products into cart and making a purchase.

This will help us counterpoise the impact of those actions, separately or combined on the time spent on the website. This suggests that we want to find out if users who both add products to the cart and make a purchase spend more time on the site than the time spent when each behavior is considered individually.

Model Without an Interaction Term

Following the model’s construction, the following outcomes were noted:

• With a mean squared error (MSE) of 2.11, the model without the interaction term accounts for roughly 80% (test R-squared) and 82% (train R-squared) of the variance in the time_spent. This indicates that time_spent predictions are, on average, 2.11 squared units off from the actual time_spent. Although this model can be improved upon, it is reasonably accurate.
• Furthermore, the plot below indicates graphically that although the model performs fairly well. There is still much room for improvement, especially in terms of capturing higher values of time_spent.

# Import libraries
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.metrics import mean_squared_error, r2_score
import statsmodels.api as sm
from sklearn.model_selection import train_test_split
import matplotlib.pyplot as plt

# Model without interaction term
X = df[['purchased', 'added_in_cart']]
y = df['time_spent']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)

# Add a constant for the intercept
X_train_const = sm.add_constant(X_train)
X_test_const = sm.add_constant(X_test)

model = sm.OLS(y_train, X_train_const).fit()
y_pred = model.predict(X_test_const)

# Calculate metrics for model without interaction term
train_r2 = model.rsquared
test_r2 = r2_score(y_test, y_pred)
mse = mean_squared_error(y_test, y_pred)

print("Model without Interaction Term:")
print('Training R-squared Score (%):', round(train_r2 * 100, 4))
print('Test R-squared Score (%):', round(test_r2 * 100, 4))
print("MSE:", round(mse, 4))
print(model.summary())


# Function to plot actual vs predicted
def plot_actual_vs_predicted(y_test, y_pred, title):

    plt.figure(figsize=(8, 4))
    plt.scatter(y_test, y_pred, edgecolors=(0, 0, 0))
    plt.plot([y_test.min(), y_test.max()], [y_test.min(), y_test.max()], 'k--', lw=2)
    plt.xlabel('Actual')
    plt.ylabel('Predicted')
    plt.title(title)
    plt.show()

# Plot without interaction term
plot_actual_vs_predicted(y_test, y_pred, 'Actual vs Predicted Time Spent (Without Interaction Term)')

Output:

Model With an Interaction Term

• A better fit for the model with the interaction term is indicated by the scatter plot with the interaction term, which displays predicted values substantially closer to the actual values.
• The model explains much more of the variance in the time_spent with the interaction term, as shown by the higher test R-squared value (from 80.36% to 90.46%).
• The model’s predictions with the interaction term are more accurate, as evidenced by the lower MSE (from 2.11 to 1.02).
• The closer alignment of the points to the diagonal line, particularly for higher values of time_spent, indicates an improved fit. The interaction term aids in expressing how user actions collectively affect the amount of time spent.

# Add interaction term
df['purchased_added_in_cart'] = df['purchased'] * df['added_in_cart']
X = df[['purchased', 'added_in_cart', 'purchased_added_in_cart']]
y = df['time_spent']
X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.3, random_state=42)

# Add a constant for the intercept
X_train_const = sm.add_constant(X_train)
X_test_const = sm.add_constant(X_test)

model_with_interaction = sm.OLS(y_train, X_train_const).fit()
y_pred_with_interaction = model_with_interaction.predict(X_test_const)

# Calculate metrics for model with interaction term
train_r2_with_interaction = model_with_interaction.rsquared
test_r2_with_interaction = r2_score(y_test, y_pred_with_interaction)
mse_with_interaction = mean_squared_error(y_test, y_pred_with_interaction)

print("nModel with Interaction Term:")
print('Training R-squared Score (%):', round(train_r2_with_interaction * 100, 4))
print('Test R-squared Score (%):', round(test_r2_with_interaction * 100, 4))
print("MSE:", round(mse_with_interaction, 4))
print(model_with_interaction.summary())


# Plot with interaction term
plot_actual_vs_predicted(y_test, y_pred_with_interaction, 'Actual vs Predicted Time Spent (With Interaction Term)')

# Print comparison
print("nComparison of Models:")
print("R-squared without Interaction Term:", round(r2_score(y_test, y_pred)*100,4))
print("R-squared with Interaction Term:", round(r2_score(y_test, y_pred_with_interaction)*100,4))
print("MSE without Interaction Term:", round(mean_squared_error(y_test, y_pred),4))
print("MSE with Interaction Term:", round(mean_squared_error(y_test, y_pred_with_interaction),4))

Output:

Comparing Model Performance

• The model predictions without the interaction term are represented by the blue points. When the actual time spent values are higher, these points are more dispersed from the diagonal line.
• The model predictions with the interaction term are represented by the red points. The model with the interaction term produces more accurate predictions. Especially for higher actual time spent values, as these points are closer to the diagonal line.

# Compare model with and without interaction term

def plot_actual_vs_predicted_combined(y_test, y_pred1, y_pred2, title1, title2):

    plt.figure(figsize=(10, 6))
    plt.scatter(y_test, y_pred1, edgecolors="blue", label=title1, alpha=0.6)
    plt.scatter(y_test, y_pred2, edgecolors="red", label=title2, alpha=0.6)
    plt.plot([y_test.min(), y_test.max()], [y_test.min(), y_test.max()], 'k--', lw=2)
    plt.xlabel('Actual')
    plt.ylabel('Predicted')
    plt.title('Actual vs Predicted User Time Spent')
    plt.legend()
    plt.show()

plot_actual_vs_predicted_combined(y_test, y_pred, y_pred_with_interaction, 'Model Without Interaction Term', 'Model With Interaction Term')

Output:

Conclusion

The improvement in the model’s performance with the interaction term demonstrates that sometimes adding interaction terms to your model may enhance its importance. This example highlights how interaction terms can capture additional information that is not apparent from the main effects alone. In practice, considering interaction terms in regression models can potentially lead to more accurate and insightful predictions.

In this blog, we first generated a synthetic dataset to simulate user behavior on an e-commerce platform. We then constructed two regression models: one without interaction terms and one with interaction terms. By comparing their performance, we demonstrated the significant impact of interaction terms on the accuracy of the model.

Key Takeaways

• Regression models with interaction terms can help to better understand the relationships between two or more variables and the target variable by capturing their combined effects.
• Including interaction terms can significantly improve model performance, as evidenced by higher R-squared values and lower MSE in this guide.
• Interaction terms are not just theoretical concepts, they can be applied to real-world scenarios.

Frequently Asked Questions

The media shown in this article is not owned by Analytics Vidhya and is used at the Author’s discretion.