Markov Inequality
27 Feb 2018Prologue To The Markov Inequality
Markov Inequality Theorem
Given that Z is a non-negative random variable, then for all t≥0, we have P(Z≥t)≤E[Z]t
proof::mjtsai1974➀let’s begin by the definition of expect value of a random variable.
E[Z]=∑(P(Z≥t)⋅{Z|Z≥t}+P(Z<t)⋅{Z|Z<t})
, where we denote {Z|Z≥t}=1,{Z|Z<t}=1
➁then:
E[Z]≥∑P(Z≥t)⋅{Z|Z≥t}…this must hold
=P(Z≥t)⋅∑{Z|Z≥t}
➂choose Qt=∑{Z|Z≥t} to be the total number of events in {Z|Z≥t}, then:
E[Z]Qt≥P(Z≥t), where Qt=0,1,2,…
➃take Qt=t could also hold to have E[Z]t≥P(Z≥t)
We just prove that P(Z≥t)≤E[Z]t