Showing posts with label Andrey Markov. Show all posts
Showing posts with label Andrey Markov. Show all posts

Saturday, August 10, 2013

Pushkin poetry, Markov chains

A Markov chain is a mathematical process that can be used to answer questions such as these:
          If the current letter I am reading is a vowel, what is the probability 
          that the next letter will be a vowel?  A consonant? 
Answers from these may be combined to create more lengthy predictions -- about the 3rd letter after a given one, or the 10th -- and so on. 
     A recent article by Brian Hayes in American Scientist (brought to my attention by Greg Coxson) alerted me to the fact that it is 100 years since the Russian mathematician A. A. Markov (1856 - 1922) announced his findings about these transition probabilities -- and, moreover, his work was based on analysis of poetry; the poetry was Eugene Onegin, a verse-novel in iambic tetrameter by Alexander Pushkin (1799-1837). Markov's analyis dealt with Pushkin's novel as a long string of alphabetic characters and he tabulated the categories of vowels and consonants for about 20,000 letters. (For a host of details, visit Hayes' careful and interesting article.)