南京邮电大学多元统计分析实验二-聚类分析

使用平台：R、Rstudio

题目描述：为了研究亚洲国家或地区的经济发展和文化教育水平，以便对亚洲国家和地区进行分类研究，进行聚类分析（要求采用多种聚类方法并比较结果的差异。）

rm(list=ls())
#install.packages("dplyr")
#install.packages("haven")
#install.packages("factoextra")
#install.packages("ggplot2")
#install.packages("dbscan")
library(haven)
library(dplyr)
library(factoextra)
library(ggplot2)
library(dbscan)
library(foreign)
library(cluster)
data <- read.spss("聚类分析（world95）.sav", to.data.frame = TRUE)
selected_features <- c("gdp_cap", "literacy", "lifeexpf")
data <- data[selected_features]
data[is.na(data)]=0
scaled_data=scale(data)
#主成分分析（PCA）进行降维处理
pca_result <- prcomp(data, scale = TRUE)

#获取主成分分析后的降维结果
reduced_data <- as.data.frame(pca_result$x)

# 使用K均值聚类进行聚类分析
# 打印聚类结果
kmeans_result <- kmeans(reduced_data, centers = 3)
print(kmeans_result)
# 密度聚类
dbscan_clusters <- dbscan(scaled_data, eps = 0.5)$cluster
print(dbscan_clusters)

# 层次聚类
hierarchical_clusters <- hclust(dist(scaled_data), method = "ward.D2")
data$Agglomerative_Cluster <- cutree(hierarchical_clusters, k = 3)
print(data$Agglomerative_Cluster)

# K均值聚类
ggplot(data, aes(x = gdp_cap, y = literacy, color = factor(KMeans_Cluster))) +
  geom_point() +
  labs(title = "KMeans Clustering", x = "GDP per Capita", y = "Literacy Rate")

# Agglomerative Clustering
ggplot(data, aes(x = gdp_cap, y = literacy, color = factor(Agglomerative_Cluster))) +
  geom_point() +
  labs(title = "Agglomerative Clustering", x = "GDP per Capita", y = "Literacy Rate")

# DBSCAN Clustering
ggplot(data, aes(x = gdp_cap, y = literacy, color = factor(DBSCAN_Cluster))) +
  geom_point() +
  labs(title = "DBSCAN Clustering", x = "GDP per Capita", y = "Literacy Rate")