A Markov chain is an instance of a so-called stochastic process, loosely speaking a random variable that evolves in time. Markov Property
Basis of Google PageRank algorithm, autocompletion.
#A First Example
- Sunny and Rainy are called the states of the chain.
- The arrows are called edges and have a transition probability associated to them.
- The sum of the probabilities of edges leaving out from a state must add up to one.
- Often represented in a Transition Matrix.
#Recurrent vs Transient
A state of a Markov chain is said to be recurrent if, starting from that state, there is a probability one of eventually returning to it.
A state is said to be transient if it is not recurrent.
#Reducible vs Irreducible
A Markov chain is said to be irreducible if it is possible to go from any state to any other state.
A Markov chain which is not irreducible is said to be reducible.
- If a Markov chain is irreducible then all its states are recurrent.
- The converse is not true!! A Markov chain whose all states are recurrent is not necessarily irreducible.
#Period of a State
The period of a state is the greatest common divisor of the length of the trips that would take to return to a state, given that you started at that state.
#Periodic vs Aperiodic
A Markov chain is aperiodic if all its states have period one. A Markov chain which is not aperiodic is said to be periodic (it has at least one state with period bigger than one).