Let . The law of total probability applied to the current markov chain state givesThe second equality holds because can only be reached from . Eliminating gives the linear recurrence relation
There are two paths from back to in two steps: and . By the markov property, their probabilities multiply along each path. Therefore
For a finite or countable partition of a set into events , the law of total probability statesThe events are pairwise disjoint and have union , so countable additivity proves the formula. Terms with contribute zero and are omitted to avoid undefined conditional probabilities.
A partition of a set is a collection of nonempty subsets whose union is and which are pairwise disjoint:For use with probabilities, the must also be events in the sample space.
For the geometric distribution, , so the probability generating function of the size-biased distribution isIf are independent and identically distributed with geometric distribution of parameter , the probability generating function of an independent sum givesAt zero, interpret the expression by its continuous extension. The probability generating functions agree, henceOne can also check the probability mass functions directly: for , there are positive integer pairs with sum , each having probability . Thus both sides assign probability to .
By the definition of a probability generating function,Termwise differentiation of the power series is valid for . At , the equality extends by the finite mean assumption, and both sides equal one.
Since , its finite mean satisfies . Every is nonnegative, andThus is a probability mass function. It defines the size-biased distribution of : an outcome of size receives times its original relative weight.
The event occurs exactly when both and . By independence,ThusThe new success probability is the chance that at least one of the two trials succeeds.
The mean is the derivative at one of the probability generating function. ThereforeEvaluation at one is valid here because the probability generating function has radius of convergence .
Put . Since takes positive integer values, . Taking differences of the given tail probabilities therefore yields, also for ,This is the probability mass function of the geometric distribution on .
The finite geometric series identity gives, for each ,For every fixed , the absolute value of this error tends to infinity as . Thusand there is no uniform convergence on .
At , . For fixed ,Thus the pointwise limit is zero everywhere on . However, with , the inequality gives , and equality occurs at . Henceso there is no uniform convergence. The moving evaluation point captures a peak missed by each fixed positive input.
Let lie in , and put and . Equality would imply . If , strict monotonicity of would give . Both are impossible, so . Therefore the inverse function is strictly increasing.
If , one of and holds. Since is strictly increasing, the corresponding values are strictly ordered, so . Thus is an injection. Every element of the image of a function equals for some , which is precisely surjectivity onto that image of a function. Hence is a bijection.
Split the power series into even and odd indices:whenever the two series converge absolutely. If , the even part is a convergent geometric series, while the odd part is bounded termwise in absolute value by , also a convergent geometric series. Hence there is absolute convergence throughout .
If , the odd terms have absolute values . Their successive ratio tends to , so these terms grow eventually and do not tend to zero. The term test for divergence therefore forces divergence of the original power series. Consequently its radius of convergence is
For fixed , the ratio of successive absolute values isThe ratio test says that a series converges absolutely when this ratio has a limit of a sequence strictly below one. Thus this power series converges absolutely for every complex ; at zero the conclusion is immediate. Its radius of convergence is .
Write . Suppose, for a contradiction, that only finitely many sign changes occur. Since every , the sign is then constant for . Set on that tail. Multiplying the recurrence relation by this constant sign givesConsequently for , and by comparison with a geometric sequence. But positivity of requires at every step, which is incompatible with . Thus the negative-parameter square-root iteration changes sign infinitely often.
Only (ii) is necessarily true. Every real satisfies , so (ii) holds automatically. For a single counterexample to (i) and (iii), choose and . The recurrence relation then gives the nonzero periodic sequenceHere , disproving (i), and , disproving (iii). This counterexample satisfies the requirement that no iterate is zero.
There are unlisted articles, also show them or only show them.