# 加载必要的库
library(arules)
library(arulesViz)
library(dplyr)

# 数据预处理
transactions_data <- read.csv("transaction_data.csv")

# 转换为事务格式
transactions <- as(split(transactions_data$product_id, 
                        transactions_data$transaction_id), "transactions")

# 频繁项集挖掘
frequent_items <- apriori(transactions, 
                         parameter = list(support = 0.01, 
                                        confidence = 0.5,
                                        minlen = 2))

# 查看规则
inspect(head(frequent_items, 10))

# 可视化关联规则
plot(frequent_items, method = "graph", 
     control = list(type = "items"))

# 用户特征工程
user_features <- user_data %>%
  mutate(
    avg_order_value = total_spent / order_count,
    days_since_last_order = as.numeric(Sys.Date() - last_order_date),
    favorite_category = most_frequent_category
  ) %>%
  select(user_id, avg_order_value, order_frequency, 
         days_since_last_order, total_spent)

# 数据标准化
scaled_features <- scale(user_features[,-1])

# K-means聚类
set.seed(123)
kmeans_result <- kmeans(scaled_features, centers = 5, nstart = 25)

# 聚类结果分析
user_clusters <- data.frame(
  user_id = user_features$user_id,
  cluster = kmeans_result$cluster
)

# 聚类特征分析
cluster_summary <- user_features %>%
  mutate(cluster = kmeans_result$cluster) %>%
  group_by(cluster) %>%
  summarise(
    count = n(),
    avg_order_value = mean(avg_order_value),
    avg_frequency = mean(order_frequency),
    avg_total_spent = mean(total_spent)
  )

# 序列数据准备
library(arulesSequences)

# 用户行为序列
user_sequences <- user_behavior %>%
  arrange(user_id, timestamp) %>%
  group_by(user_id) %>%
  summarise(
    sequence = paste(action_type, collapse = " -> "),
    .groups = 'drop'
  )

# 序列模式挖掘
sequences <- as(user_sequences$sequence, "transactions")
seq_patterns <- cspade(sequences, 
                      parameter = list(support = 0.05),
                      control = list(verbose = FALSE))

# 分析高频序列模式
frequent_patterns <- as(seq_patterns, "data.frame")
head(frequent_patterns[order(-frequent_patterns$support), ])