问spmf中关联规则中的r子集
EN

Stack Overflow用户

提问于 2014-08-01 00:36:28

回答 1查看 477关注 0票数 1

代码：

data=read.csv("./spmf1234.csv",header=F);

df<- data.frame(do.call('rbind', strsplit(as.character(data$V1),'#',fixed=F)));

df2<- data.frame(do.call('rbind', strsplit(as.character(df$X1),'==>',fixed=F)));

df3=cbind(df2,df);

colnames(df3)=c("lhs","rhs","rule","support","confidence","lift");

df4 <- subset(df3, select = c(lhs,rhs,support,confidence,lift));

final=subset(df4,lhs!=1);

我在csv文件上尝试上面的代码，我已经从spmf中获得了使用from来获得关联规则。我希望所有的规则'1‘在lhs被删除，但这是无效的。

csv文件：

2 ==> 1 #SUP: 1 #CONF: 0.33333 #升降: 0.66667

1 ==> 2 #SUP: 1 #CONF: 0.33333 #升降: 0.66667

3 ==> 1 #SUP: 2 #CONF: 0.5 #升降:1

1 ==> 3 #SUP: 2 #CONF: 0.66667 #升降:1

3 ==> 2 #SUP: 2 #CONF: 0.5 #升降:1

2 ==> 3 #SUP: 2 #CONF: 0.66667 #升降机:1

2 3 ==> 1 #SUP: 1 #CONF: 0.5 #升降:1

1 3 ==> 2 #SUP: 1 #CONF: 0.5 #升降:1

1 2 ==> 3 #SUP: 1 #CONF: 1 #LIFT: 1.5

3 ==> 1 2 #SUP: 1 #CONF: 0.25 #升力: 1.5

2 ==> 1 3 #SUP: 1 #CONF: 0.33333 #LIFT: 1

1 ==> 2 3 #SUP: 1 #CONF: 0.33333 #LIFT: 1

subset

apriori

人脸识别特惠活动

精准实时的人脸检测、分析、识别等服务，在线、离线满足多样需求，可应用于智慧零售、智慧社区、在线娱乐等多种应用场景

回答 1

Stack Overflow用户

回答已采纳

发布于 2014-08-01 00:44:13

final=subset(df4,lhs!=1)

将字符与不起作用的数字进行比较：

# > as.character(df4$lhs)
# [1] "2 "   "1 "   "3 "   "1 "   "3 "   "2 "   "2 3 " "1 3 " "1 2 " "3 "   "2 "   "1 "

您可能需要使用一个正则表达式，如insead：

final = subset(df4, !grepl("1\\b", lhs))
# > final
#    lhs   rhs support     confidence          lift
# 1    2     1  SUP: 1  CONF: 0.33333  LIFT: 0.66667
# 3    3     1  SUP: 2      CONF: 0.5        LIFT: 1
# 5    3     2  SUP: 2      CONF: 0.5        LIFT: 1
# 6    2     3  SUP: 2  CONF: 0.66667        LIFT: 1
# 7  2 3     1  SUP: 1      CONF: 0.5        LIFT: 1
# 10   3   1 2  SUP: 1     CONF: 0.25      LIFT: 1.5
# 11   2   1 3  SUP: 1  CONF: 0.33333        LIFT: 1

Add:

## data preperation
data <- readLines(con = textConnection("  
2 ==> 1 #SUP: 1 #CONF: 0.33333 #LIFT: 0.66667
1 ==> 2 #SUP: 1 #CONF: 0.33333 #LIFT: 0.66667
3 ==> 1 #SUP: 2 #CONF: 0.5 #LIFT: 1
1 ==> 3 #SUP: 2 #CONF: 0.66667 #LIFT: 1
3 ==> 2 #SUP: 2 #CONF: 0.5 #LIFT: 1
2 ==> 3 #SUP: 2 #CONF: 0.66667 #LIFT: 1
2 3 ==> 1 #SUP: 1 #CONF: 0.5 #LIFT: 1
1 3 ==> 2 #SUP: 1 #CONF: 0.5 #LIFT: 1
1 2 ==> 3 #SUP: 1 #CONF: 1 #LIFT: 1.5
3 ==> 1 2 #SUP: 1 #CONF: 0.25 #LIFT: 1.5
2 ==> 1 3 #SUP: 1 #CONF: 0.33333 #LIFT: 1
1 ==> 2 3 #SUP: 1 #CONF: 0.33333 #LIFT: 1"))
r <- regexec(pattern = "([0-9 ]+)\\s==>\\s([0-9 ]+)\\s\\#SUP:\\s([0-9.]+)\\s\\#CONF:\\s([0-9.]+)\\s\\#LIFT:\\s([0-9.]+)", 
             text = data)
m <- regmatches(data, r)
df <- setNames(as.data.frame(do.call(rbind, lapply(m, "[", -1)), stringsAsFactors = FALSE), 
               c("lhs", "rhs", "support", "confidence", "lift"))

## rows to include/exclude
include <- sapply(strsplit(df$lhs, " "), function(x) !any(as.integer(x) %in% 1:1000))
df[include, ]