IAI Actuarial Core Principles · Risk Modelling and Survival Analysis · Markov processes
A discrete-time Markov chain on states {1,2,3,4} has transition matrix rows: state 1: (0.5, 0.5, 0, 0); state 2: (0.4, 0.6, 0, 0); state 3: (0, 0, 0.3, 0.7); state 4: (0.2, 0.2, 0.3, 0.3). Which statement about the classes of this chain is correct?
{1,2} is closed because no transitions leave it, while states 3 and 4 communicate with each other but can leak into {1,2}, so their class is not closed and those states are transient.
- AThe chain is irreducible because every state has some outgoing transition
- B{1,2} is a closed class and {3,4} is also a closed class
- C{1,2} is a closed class, and states 3 and 4 are in a single non-closed communicating classCorrect
- DStates 3 and 4 are absorbing states
- {1,2} is not closed because state 4 leads into it
Explanation
From states 1 and 2 the chain can only reach 1 and 2, so {1,2} is closed. State 3 reaches 4 and state 4 reaches 3, so they communicate. State 4 can also move to 1 or 2 and never returns, so {3,4} is not closed. Option 4 confuses entry into a class with leaving it.
Did you get it right without looking?
One question tells you little. A timed set on Markov processes shows your real accuracy, how long you take and where you lose marks.
More Markov processes questions
- A two-state process (Alive A, Dead D) has constant force of mortality 0.04 per year, with no other transitions. What is the probability that…
- A healthy-sick two-state model has constant rates: healthy to sick 0.2 per year and sick to healthy 0.6 per year. The process is in the heal…
- For a time-homogeneous Markov jump process with transition probabilities p_ij(t) and transition rates mu_ij, the Kolmogorov forward equation…
- A two-state Markov jump process has rate a from state 0 to 1 and rate b from 1 to 0, with a=0.3 and b=0.5. Starting in state 0, p_00(t) take…
- In a time-homogeneous Markov jump process, the jump chain has transition probability from state i to j equal to:
- In a time-homogeneous continuous-time Markov jump process, the holding time in state i has which distribution, and what is its parameter?