Create a markov chain object from transition matrix

Question

I have a transition matrix with 4 rows and 16 columns, containing the probability to transition from one Wikipedia article to another. The sum of my rows have been normalised to equal 1.

Here is an overview:

dput(head(res,4))
structure(c(0.0269201326869099, 0.221418475097697, 0.0572200491955321, 
0.0234890665263753, 0.0931927078903348, 0.203375737923007, 0.310133473123916, 
0.220087000151008, 0.00266507896005217, 0.0293922008813503, 0.00467760796806323, 
0.0045399854837374, 0.00598225170820221, 0.0235719630830631, 
0.0370176216782935, 0.00138830028204416, 0.0424711519378526, 
0.0206618441839195, 0.0414532844066293, 0.0344054908493962, 0.0334552465198038, 
0.0453978548266401, 0.166498649139078, 0.0243074330084224, 0.391312976666383, 
0.34181425126798, 0.131335940965361, 0.113533735696854, 0.190425562076493, 
0.0283528727030847, 0.00129037461187951, 0.0467297003707005, 
0.0281959683592753, 0.061278789390538, 0, 0.0299483162596755, 
0, 0.0158809345638979, 0.0305254244122747, 0, 0.0255592413030535, 
0, 0.0753659421750877, 0.0176874327161486, 0, 0, 0.0458082987217227, 
0, 0.0268350769754189, 0, 0.0630670591556111, 0.260016464754222, 
0, 0, 0.000846808339045929, 0.00347318631964031, 0.0287771823877973, 
0.00253596075496799, 0.0145570385902657, 0.0108920681777219, 
0.104207422528423, 0.00631911532385466, 0.0202024275172386, 0.209501819404054
), .Dim = c(4L, 16L), .Dimnames = list(c("Selected Disease Article", 
"Selected Food Scandal Article", "Selected General Article", 
"Selected Pathogen Article"), c("Animal health", "Biology", "Environment", 
"Food industry", "General", "Human and animal health", "Human health", 
"Medicine", "Other", "Related to a Food scandal", "Related to Antibiotic resistance", 
"Related to Food poisoning", "Selected Disease Article", "Selected Food Scandal Article", 
"Selected General Article", "Selected Pathogen Article")))

I followed the code below but I get stuck on the definition of my markov chain object

states <- as.character(1:16)

mc <- new(
    "markovchain",
    states = states,
    byrow = TRUE,
    transitionMatrix = res,
    name = "random_walk");

I get the error: Error in dimnames(x) <- dn : length of 'dimnames' [2] not equal to array extent

What is the problem in my code? Am I doing something wrong in the creation of my markov chain function?

Please edit your post to make a minimally working example: https://stackoverflow.com/questions/5963269/how-to-make-a-great-r-reproducible-example — emilliman5, Apr 05 '19 at 13:05

score 0 · Answer 1 · answered Apr 05 '19 at 14:30

This is more of a fundamental issue with your (Transition) Markov Matrix which is a m x n matrix, it should be a n x n matrix:

http://www.math.harvard.edu/~knill/teaching/math19b_2011/handouts/lecture33.pdf

This is likely why you receive the dimnames error as the object expects the dimnames to be of same length, see page 4 of the docs:

https://cran.r-project.org/web/packages/markovchain/markovchain.pdf

Cool viz of markov chains:

http://setosa.io/ev/markov-chains/

Create a markov chain object from transition matrix

1 Answers1