R:Montecarlo 使用转移矩阵模拟

Question

我有一个转换矩阵：

a
           A          B         C     D         E          F          G          H         I
A 0.00000000 0.66666667 0.0000000 0.000 0.0000000 0.00000000 0.00000000 0.33333333 0.0000000
B 0.08823529 0.02941176 0.2941176 0.000 0.2352941 0.05882353 0.02941176 0.11764706 0.1470588
C 0.00000000 0.37500000 0.0000000 0.000 0.4166667 0.00000000 0.00000000 0.08333333 0.1250000
D 0.00000000 0.00000000 0.3333333 0.000 0.0000000 0.00000000 0.33333333 0.33333333 0.0000000
E 0.00000000 0.50000000 0.2307692 0.000 0.0000000 0.00000000 0.00000000 0.07692308 0.1923077
F 0.00000000 0.00000000 0.0000000 0.000 0.5000000 0.00000000 0.00000000 0.00000000 0.5000000
G 0.00000000 0.50000000 0.0000000 0.500 0.0000000 0.00000000 0.00000000 0.00000000 0.0000000
H 0.00000000 0.27272727 0.3636364 0.000 0.1818182 0.00000000 0.00000000 0.00000000 0.1818182
I 0.00000000 0.31250000 0.1875000 0.125 0.3125000 0.00000000 0.00000000 0.06250000 0.0000000

我的数据集中有一个分类变量：

state=c("G" ,"I" ,"G", "C", "D", "I","A" ,"G", "G" ,"H", "C", "D" ,"C", "H" "F", "B", "F" ,"G" ,"D", "E" ,"B" ,"H" ,"E" ,"C" ,"F" ,"H", "C", "H" ,"F" ,"H")

现在我想使用我的转换矩阵来模拟一个变量，就像我的变量 state 使用 Monte carlo 方法一样。你能告诉我哪个 R 包或函数可以帮助我做我的模拟吗！

Answer 1

您可以使用包 markovchain 中的 markovchainSequence。示例：

mat <- as.matrix(read.table(text="
0.00000000 0.66666667 0.0000000 0.000 0.0000000 0.00000000 0.00000000 0.33333333 0.0000000
0.08823529 0.02941176 0.2941176 0.000 0.2352941 0.05882353 0.02941176 0.11764706 0.1470588
0.00000000 0.37500000 0.0000000 0.000 0.4166667 0.00000000 0.00000000 0.08333333 0.1250000
0.00000000 0.00000000 0.3333333 0.000 0.0000000 0.00000000 0.33333333 0.33333333 0.0000000
0.00000000 0.50000000 0.2307692 0.000 0.0000000 0.00000000 0.00000000 0.07692308 0.1923077
0.00000000 0.00000000 0.0000000 0.000 0.5000000 0.00000000 0.00000000 0.00000000 0.5000000
0.00000000 0.50000000 0.0000000 0.500 0.0000000 0.00000000 0.00000000 0.00000000 0.0000000
0.00000000 0.27272727 0.3636364 0.000 0.1818182 0.00000000 0.00000000 0.00000000 0.1818182
0.00000000 0.31250000 0.1875000 0.125 0.3125000 0.00000000 0.00000000 0.06250000 0.000000",
           header=FALSE,stringsAsFactors = FALSE))
rownames(mat) <- colnames(mat) <- LETTERS[1:9]
mat[,9] <- 1-rowSums(mat[,1:8]) #To make sure your rows sum to 1

statesNames <- LETTERS[1:9]
markovchain_object <- new("markovchain", states = statesNames, transitionMatrix = mat)

markovchainSequence(n=10, markovchain = markovchain_object)

 [1] "C" "H" "C" "E" "C" "B" "C" "E" "B" "G"

R:Montecarlo 使用转移矩阵模拟

R:Montecarlo simulation using transition matrix

r

simulation

montecarlo