R语言做贝叶斯网络结构学习

本文使用pcalg包做贝叶斯网络结构的学习

实例分析

pcalg包的安装

安装pcalg包时由于有些依赖包已经被CRAN撤掉，所以安装起来有些技巧。
以下是安装代码：

install.packages("pcalg")
## 如果https://不行可以改成http://试试
source("https://bioconductor.org/biocLite.R")
#如果遇到Update all/some/none? [a/s/n]:  填a就可以了
biocLite("graph") 
biocLite("RBGL")
biocLite("Rgraphviz")

贝叶斯网络结构学习

iris数据集

我们使用iris数据集进行学习，这是一个针对连续变量的结构学习的例子：

suffStat <-list(C = cor(iris[,1:4]), n = nrow(iris))
pc.fit <- pc(suffStat, indepTest = gaussCItest,
             alpha = 0.01,labels =names(iris[1:4]),verbose = TRUE)
pc.fit
plot(pc.fit)

Adult数据集

该数据从美国1994年人口普查数据库抽取而来，可以用来预测居民收入是否超过50K/year。该数据集类变量为年收入是否超过50k$，属性变量包含年龄，工种，学历，职业，人种等重要信息，值得一提的是，14个属性变量中有7个类别型变量。

• age – The age of the individual
• type_employer – The type of employer the individual has. Whether they are government, military, private, an d so on.
• fnlwgt – The # of people the census takers believe that observation - represents. We will be ignoring this variable
• education – The highest level of education achieved for that individual
• education_num – Highest level of education in numerical form
• marital – Marital status of the individual
• occupation – The occupation of the individual
• relationship – A bit more difficult to explain. Contains family relationship values like husband, father, and so on, but only contains one per observation. I’m not sure what this is supposed to represent
• race – descriptions of the individuals race. Black, White, Eskimo, and so on
• sex – Biological Sex
• capital_gain – Capital gains recorded
• capital_loss – Capital Losses recorded
• hr_per_week – Hours worked per week
• country – Country of origin for person
• income – Boolean Variable. Whether or not the person makes more than $50,000 per annum income.

对于离散数据，我们使用 G2 统计量来检验条件独立性。
接来下探索一下，教育，婚姻状态，职业，收入的关系是怎样的：

url <- "http://archive.ics.uci.edu/ml/machine-learning-databases/adult/adult.data"
adult <- read.csv(url, strip.white = TRUE, header = FALSE,
    col.names=c("age", "type_employer", "fnlwgt", "education", 
    "education_num","marital", "occupation", "relationship", "race","sex",
    "capital_gain", "capital_loss", "hr_per_week","country", "income"))

> adult<-adult[,c(4,6,7,15)] 
> head(adult)
  education            marital        occupation income
1 Bachelors      Never-married      Adm-clerical  <=50K
2 Bachelors Married-civ-spouse   Exec-managerial  <=50K
3   HS-grad           Divorced Handlers-cleaners  <=50K
4      11th Married-civ-spouse Handlers-cleaners  <=50K
5 Bachelors Married-civ-spouse    Prof-specialty  <=50K
6   Masters Married-civ-spouse   Exec-managerial  <=50K

adult2<-sapply(adult,function(x){
  temp<-as.integer(factor(x))-1
})
nlev<-sapply(adult,function(x){
  temp<-factor(x)
  length(levels(temp))
})
adult2<-as.data.frame(adult2)
suffStat <- list(dm = adult2, nlev = nlev, adaptDF = FALSE)

pc.fit <- pc(suffStat, indepTest = disCItest,
             alpha = 0.01,labels =names(adult2),verbose = TRUE)
pc.fit

plot(pc.fit)

该包的详细说明：

• Causal Inference using Graphical Models with the Package pcalg

• Package ‘pcalg’

作为分享主义者(sharism)，本人所有互联网发布的图文均遵从CC版权，转载请保留作者信息并注明作者a358463121专栏:http://blog.csdn.net/a358463121，如果涉及源代码请注明GitHub地址：https://github.com/358463121/。商业使用请联系作者。