svm

# Load necessary library
library(ggplot2)
library(e1071)  # For SVM

# Set random seed for reproducibility
set.seed(123)

# Generate synthetic data
n <- 100
dosage <- runif(n, 0, 1)  # Dosages between 0 and 1
cured <- ifelse(dosage > 0.3 & dosage < 0.7, 1, 0)  # Cured in the "just right" range

# Create a data frame
data <- data.frame(dosage = dosage, cured = as.factor(cured))

# Plot the original data
ggplot(data, aes(x = dosage, y = cured, color = cured)) +
  geom_point(size = 3) +
  labs(title = "Drug Dosage vs. Cure Status",
       x = "Dosage",
       y = "Cure Status (0 = Not Cured, 1 = Cured)") +
  scale_color_manual(values = c("red", "green")) +
  theme_minimal()

# Transform data: Add dosage^2 as a new dimension
data$dosage_sq <- data$dosage^2

# Train a support vector machine
svm_model <- svm(cured ~ dosage + dosage_sq, data = data, kernel = "linear", cost = 1)

# Generate a grid of points to visualize decision boundary
grid <- expand.grid(
  dosage = seq(0, 1, length.out = 100),
  dosage_sq = seq(0, 1, length.out = 100)
)

# Predict classifications on the grid
grid$cured <- predict(svm_model, grid)

# Plot the transformed data and decision boundary
ggplot(data, aes(x = dosage, y = dosage_sq, color = cured)) +
  geom_point(size = 3) +
  geom_contour(data = grid, aes(x = dosage, y = dosage_sq, z = as.numeric(cured)), 
               breaks = 0.5, color = "blue") +
  labs(title = "Support Vector Classifier in Transformed Space",
       x = "Dosage",
       y = "Dosage^2") +
  scale_color_manual(values = c("red", "green")) +
  theme_minimal()

## Warning: `stat_contour()`: Zero contours were generated

## Warning in min(x): no non-missing arguments to min; returning Inf

## Warning in max(x): no non-missing arguments to max; returning -Inf