1、导入数据

#2018年物料数据
library(xlsx)
x=read.xlsx(“material_sales_order.xlsx”,sheetIndex=1,encoding = “UTF-8”)
x=read.xlsx(“material_sales_order_动态.xlsx”,sheetIndex=1,encoding = “UTF-8”)
x=x[xKaTeX parse error: Expected 'EOF', got '#' at position 17: …rder_nums>5,] #̲考虑数据的实际情况，将数据值为…km.cluster<=2,] #选择聚类的1和2类

2、观测数据分布

#物料订单额数据
boxplot(x $order_amount) qqnorm(x$order_amount)
#物料订单数数据
boxplot(xKaTeX parse error: Expected 'EOF', got '#' at position 13: order_nums) #̲数据的正态分布检验 qqnor…order_nums)

3、对原始数据进行标准化数据

X=cbind(x,scale(x $order_amount),scale(x$order_nums)) #对数据做标准化
write.xlsx(X,“material_sales_ordery_标准化.xlsx”)

4、进行聚类分析

##k-means聚类确定聚类个数
d=data.frame(x $order_nums) mydata <- d wss <- (nrow(mydata)-1)*sum(apply(mydata,2,var)) for (i in 2:15) wss[i] <- sum(kmeans(mydata,centers=i)$withinss)
###这里的wss(within-cluster sum of squares)是组内平方和
plot(1:15, wss, type=“b”, xlab=“聚类个数”,ylab=“误差(平方和)”)
##动态聚类
km=kmeans(x $order_nums,6) plot(x$order_nums, col = km $cluster, pch =1, xlab="序号",ylab="订单数量") X=cbind(x,km$cluster) #读取聚类结果
write.xlsx(X,“material_sales_order_动态.xlsx”)
#自动计算的订单数散布最小的聚类个数
library(ykmeans)
a=data.frame(x $order_nums) km=ykmeans(a,"x.order_nums","x.order_nums",3:6) table(km$cluster)

考虑最近一次订单发生的时间
是否可以考虑使用因子分析对其进行评分