Logistic Regression for Binary Classification and Visualization

# File Name: logistic_regression_classification.R

# Author: Gerard King - www.gerardking.dev

# Title: Logistic Regression for Binary Classification and Visualization

# Description: This R program generates synthetic data for a binary classification problem, 

# fits a logistic regression model to the data, and visualizes the decision boundary.

# It demonstrates the use of logistic regression for predicting binary outcomes.

# Use Cases:

# - Predicting binary outcomes (e.g., yes/no, 0/1)

# - Classifying data points into two categories

# - Solving classification problems in various domains (e.g., marketing, medical diagnostics)

# Audience:

# - Data scientists and machine learning practitioners

# - Students learning about classification techniques

# - Researchers working on binary classification problems

# Blue Team Uses:

# - Identifying whether a security event is benign or suspicious

# - Classifying system logs into different categories (e.g., normal vs. attack)

# - Predicting binary outcomes like whether a system will experience a breach or not

# Red Team Uses:

# - Using logistic regression to classify attack attempts based on different features

# - Predicting the likelihood of a successful attack based on various factors

# - Analyzing the likelihood of detecting a breach given certain system metrics

# Current Date: 2025-03-06

# Load necessary library

library(ggplot2)

# Set seed for reproducibility

set.seed(404)

# Generate synthetic dataset with 2 features (x1, x2) and binary outcome (y)

n <- 200

x1 <- rnorm(n)

x2 <- rnorm(n)

y <- ifelse(0.5 * x1 + 0.5 * x2 + rnorm(n) > 0, 1, 0)  # Binary outcome

# Create a data frame

data <- data.frame(x1 = x1, x2 = x2, y = as.factor(y))

# Fit a logistic regression model

logit_model <- glm(y ~ x1 + x2, data = data, family = binomial)

# Make predictions using the logistic regression model

data$predicted <- predict(logit_model, type = "response")

data$predicted_class <- ifelse(data$predicted > 0.5, 1, 0)

# Create a scatter plot with actual and predicted classes

ggplot(data, aes(x = x1, y = x2, color = as.factor(predicted_class))) +

  geom_point(size = 3) +

  labs(title = "Logistic Regression Classification", 

       x = "Feature 1 (x1)", y = "Feature 2 (x2)", color = "Predicted Class") +

  scale_color_manual(values = c("red", "blue")) +

  theme_minimal()

# Plot the decision boundary

ggplot(data, aes(x = x1, y = x2)) +

  geom_point(aes(color = as.factor(predicted_class)), size = 3) +

  stat_contour(data = data.frame(x1 = seq(min(x1), max(x1), length.out = 100),

                                 x2 = seq(min(x2), max(x2), length.out = 100)),

               aes(x = x1, y = x2, z = predict(logit_model, newdata = data.frame(x1 = x1, x2 = x2), type = "response")),

               color = "black", bins = 1) +

  labs(title = "Decision Boundary of Logistic Regression", 

       x = "Feature 1 (x1)", y = "Feature 2 (x2)") +

  theme_minimal()

# Display the model summary

cat("Logistic Regression Model Summary:\n")

summary(logit_model)

# Print the current date for reference

cat("Date of execution:", Sys.Date(), "\n")