Example homework- Inferring a Markov process from data

Frank Ritter

DRAFT of 24 Oct 00 DRAFT

Part 1. Markov processes

Read about Markov processes in Washburn et al. or on the net (baseball examples) or here based on what I write.

They are the idea that tokens are generated without reference to history but are based on the state that the system is in.



Part 2. Compute a Markov process from data

Given the data available through this LINK, generate the minimum Markov process that generates the data.

After you have done so, answer the following questions:

(a) what does it mean to be a minimum process?

(b) what would you do if there was one token that did not match? Would you pitch your proces, or note it as an exception?

(c) what could you use this Markov process for?



DATA would be something like a bunch of key presses on a phone, a path though the city centre of State College.