R:如何使用分类特征和数字特征来度量差异

data <- data.frame(
  category = c("A", "B", "A", "C", "B"),
  number = c(10, 15, 20, 25, 30)
)

install.packages("caret")
library(caret)

# 将数据拆分为训练集和测试集
set.seed(123)
trainIndex <- createDataPartition(data$category, p = 0.8, list = FALSE)
trainData <- data[trainIndex, ]
testData <- data[-trainIndex, ]

# 使用训练集拟合一个朴素贝叶斯分类器
nbModel <- train(category ~ ., data = trainData, method = "nb")

# 使用测试集进行预测
predictions <- predict(nbModel, testData)

# 计算混淆矩阵
confusionMatrix(predictions, testData$category)

install.packages("philentropy")
library(philentropy)

# 提取数字特征
numberFeature <- data$number

# 计算欧氏距离矩阵
euclideanDistances <- dist(numberFeature, method = "euclidean")

install.packages("class")
library(class)

# 将分类特征转换为数值型
data$categoryNumeric <- as.numeric(factor(data$category))

# 合并分类特征和数字特征
combinedFeatures <- data[, c("categoryNumeric", "number")]

# 将数据拆分为训练集和测试集
set.seed(123)
trainIndex <- createDataPartition(combinedFeatures$categoryNumeric, p = 0.8, list = FALSE)
trainData <- combinedFeatures[trainIndex, ]
testData <- combinedFeatures[-trainIndex, ]

# 使用k-NN算法进行分类
predictions <- knn(train = trainData, test = testData, cl = data$category[trainIndex], k = 3)

# 计算准确率
mean(predictions == data$category[-trainIndex])