#title Sequence Clustering {{{ x <- read.table("clipboard", header=T) head(x) > head(x) acnt_key X2013.09 X2013.10 X2013.11 X2013.12 X2014.01 X2014.02 1 48935 vvip vvip vip 일반 일반 일반 2 117140 vvip X X X X 일반 3 252239 vvip X X X X X 4 616962 vvip X X X X X 5 667645 vvip vip vip 일반 vip X 6 1254180 vvip X X X X X #library(TraMineR) #x.labels <- c('vvip','vip','BU','x') #x.scode <- c('vvip','vip','BU','x') #x.seq <- seqdef(x, 2:7, states = x.scode, labels = x.labels) x.seq <- seqdef(x, c(2:7)) head(x.seq) #seqiplot(x.seq) #seqHtplot(x.seq) #seqfplot(x.seq) seqdplot(x.seq, border = NA) #library(cluster) submat <- seqsubm(x.seq, method = "TRATE") dist.om1 <- seqdist(x.seq, method = "OM", indel = 1, sm = submat) clusterward1 <- agnes(dist.om1, diss = TRUE, method = "ward") #plot(clusterward1) cl1.4 <- cutree(clusterward1, k = 4) cl1.4fac <- factor(cl1.4, labels = paste("Type", 1:4)) #seqdplot(x.seq, group = cl1.4fac) seqfplot(x.seq, group = cl1.4fac, border = NA) mvad.seqe <- seqecreate(x.seq) fsubseq <- seqefsub(mvad.seqe, pMinSupport = 0.05) plot(fsubseq[1:10], cex=0.6, col=terrain.colors(10)) discr <- seqecmpgroup(fsubseq, group = cl1.4fac) plot(discr[1:6]) }}} {{{ 구매패턴 Support Count ---------------------------------- ---------- ----- (VVIP)-(VVIP>일반)-(일반>구매안함) 0.31372549 80 아래와 같은 의미들.. VVIP,일반,구매안함,구매안함,구매안함,구매안함 VVIP,일반,구매안함,구매안함,구매안함,일반 VVIP,일반,일반,구매안함,구매안함,구매안함 VVIP,일반,일반,VVIP,일반,구매안함 VVIP,VVIP,VVIP,VVIP,일반,구매안함 VVIP,VVIP,일반,일반,구매안함,구매안함 }}} ==== 참고 자료 ==== * http://mephisto.unige.ch/pub/TraMineR/doc/TraMineR-Users-Guide.pdf * http://www.bristol.ac.uk/cmm/software/support/workshops/materials/sequence.pdf --> 이거 좋음