Question

我正在使用R包点击流来估算转换概率。

输出如下：

Higher-Order Markov Chain (order=2)

Transition Probabilities:

Lag:  1 
lambda:  1 
                                   A Conversion                        D
A          0,25999823819591261209183          0 0,0605291772381297554784
Conversion 0,02034883720930232703372          0 0,0021747009786154403166
D          0,00162966878083157162491          0 0,6451612903225806272545
E          0,00000000000000000000000          0 0,0000000000000000000000
F          0,00132135306553911205212          0 0,0007249003262051467722
N          0,00603417899929527875441          0 0,0181225081551286693049
Null       0,62182875264270609516615          1 0,1808626313881841196629
O          0,00792811839323467187901          0 0,0094237042406669080385
P          0,00017618040873854828084          0 0,0000000000000000000000
Start      0,00000000000000000000000          0 0,0000000000000000000000
T          0,08073467230443974462784          0 0,0830010873504893054164
                                 E                        F                         N Null
A          0,035714285714285712303 0,0675675675675675713183 0,01131221719457013545618    0
Conversion 0,071428571428571424606 0,0405405405405405427910 0,00061703002879473468293    0
D          0,000000000000000000000 0,0022522522522522522327 0,00102838338132455776874    0
E          0,392857142857142849213 0,0000000000000000000000 0,00000000000000000000000    0
F          0,000000000000000000000 0,3355855855855855995884 0,00123406005758946936586    0
N          0,000000000000000000000 0,0090090090090090089309 0,09543397778691896371495    0
Null       0,214285714285714273819 0,4797297297297297147267 0,86960098724804602010607    0
O          0,000000000000000000000 0,0045045045045045044654 0,00143973673385438096298    0
P          0,000000000000000000000 0,0000000000000000000000 0,00020567667626491157001    0
Start      0,000000000000000000000 0,0000000000000000000000 0,00000000000000000000000    0
T          0,285714285714285698425 0,0608108108108108141865 0,01912793089263677484557    0
                                   O                       P                     Start
A          0,05175117616309461709978 0,059999999999999997780 0,23964753542773373218289
Conversion 0,01568217459487715459421 0,020000000000000000416 0,00000000000000000000000
D          0,00209095661265028742379 0,000000000000000000000 0,01343022430996705478268
E          0,00026136957658128592797 0,000000000000000000000 0,00022068443701823799359
F          0,00156821745948771567626 0,000000000000000000000 0,00345213512192815144206
N          0,00836382645060114969515 0,020000000000000000416 0,06317880168350699376933
Null       0,62859383167799265379472 0,520000000000000017764 0,00000000000000000000000
O          0,20987976999477261919047 0,000000000000000000000 0,04090543671873768383218
P          0,00000000000000000000000 0,320000000000000006661 0,00045713204810920727824
Start      0,00000000000000000000000 0,000000000000000000000 0,00000000000000000000000
T          0,08180867746994249789516 0,059999999999999997780 0,63870805025299892676571
                                   T
A          0,01751190166393681163504
Conversion 0,00991932891453763294909
D          0,00116337808256922864945
E          0,00003061521269919022833
F          0,00048984340318704365329
N          0,00254106265403278904630
Null       0,62369311310790331859266
O          0,00326052015246375932397
P          0,00000000000000000000000
Start      0,00000000000000000000000
T          0,34139023680867025545282

Lag:  2 
lambda:  0 
                                   A Conversion                         D
A          0,31341719077568136464507          0 0,05353982300884956024856
Conversion 0,02154670393664104288689          0 0,00353982300884955747300
D          0,00384346610761705093084          0 0,63628318584070797658825
E          0,00000000000000000000000          0 0,00000000000000000000000
F          0,00279524807826694605267          0 0,00044247787610619468412
N          0,00815280689494526054906          0 0,01548672566371681394437
Null       0,48742138364779874493138          0 0,18008849557522124684716
O          0,01059864896342883715252          0 0,01061946902654867241900
P          0,00058234334963894716659          0 0,00000000000000000000000
Start      0,00000000000000000000000          0 0,00000000000000000000000
T          0,15164220824598181880383          0 0,10000000000000000555112
...

现在我所做的是将输出写入文本文件，然后以矩阵形式读取

sink("mc_output.txt")
mc
sink()


# Lag 1
mydat11 = read.table("mc_output.txt", skip = 6, nrows = 11)
mydat12 = read.table("mc_output.txt", skip = 18, nrows = 11)
mydat13 = read.table("mc_output.txt", skip = 30, nrows = 11)
mydat14 = read.table("mc_output.txt", skip = 42, nrows = 11)
...

当矩阵的尺寸发生变化时，这种方法非常容易出错。想法是直接从内存中读取输出。有没有人有任何聪明的解决方案如何实现呢？

Answer 1

给出

library(clickstream)
clickstreams <- c("User1,h,c,c,p,c,h,c,p,p,c,p,p,o",
"User2,i,c,i,c,c,c,d",
"User3,h,i,c,i,c,p,c,c,p,c,c,i,d",
"User4,c,c,p,c,d",
"User5,h,c,c,p,p,c,p,p,p,i,p,o",
"User6,i,h,c,c,p,p,c,p,c,d")
csf <- tempfile()
writeLines(clickstreams, csf)
cls <- readClickstreams(csf, header = TRUE)
mc <- fitMarkovChain(cls)

您可以使用

访问转换矩阵

mc@transitions[[1]]
#      c d   h         i o          p
# c 0.32 0 0.8 0.5714286 0 0.47058824
# d 0.12 0 0.0 0.1428571 0 0.00000000
# h 0.04 0 0.0 0.1428571 0 0.00000000
# i 0.12 0 0.2 0.0000000 0 0.05882353
# o 0.00 0 0.0 0.0000000 0 0.11764706
# p 0.40 0 0.0 0.1428571 0 0.35294118

mc对象也有其他一些插槽：

slotNames(mc)
# [1] "states"                 "order"                  "transitions"            "lambda"                
# [5] "logLikelihood"          "observations"           "start"                  "end"                   
# [9] "transientStates"        "absorbingStates"        "absorbingProbabilities"

将R输出从内存读入矩阵（包点击流）

1 个答案: