This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: Statistics 612: L p spaces, metrics on spaces of probabilites, and connections to estimation Moulinath Banerjee March 17, 2009 1 L p spaces and Hilbert spaces We first formally define L p spaces. Consider a measure space ( X , A , μ ), where μ is a ( σfinite) measure. Let F be the set of all real–valued measurable functions defined on X . The space L p ( μ ) comprises that subset of functions of F that have finite p ’th moments (here p ≥ 1). In other words: L p ( μ ) = { f ∈ F : Z  f ( x )  p d μ ( x ) < ∞} . When the measure μ is a probability P , we obtain the class of all random variables on the probability space, ( X , A , P ) that have finite p ’th moments. This is a normed linear space over R with the norm (length) of the vector f (this is called the L p norm) being given by f p = Z  f ( x )  p d μ ( x ) ¶ 1 /p . The above norm induces a metric d where d ( f, g ) = f g p . Note that d ( f, g ) = 0 if and only if f = g a.e. μ , in which case we identify f with g . The L p norm, like all worthy norms, satisfies the triangle inequality: f + g p ≤ f p + g p ; this is precisely Minkowski’s inequality. For random variables X, Y defined on the same probability space and having finite p ’th moments, Minkowski’s inequality states: E (  X + Y  p ) 1 /p ≤ E (  X  p ) 1 /p + E (  Y  p ) 1 /p . Minkowski’s inequality is a consequence of H¨older’s inequality which states that for measurable realvalued functions f, g defined on X , we have:  Z f ( x ) g ( x ) d μ ( x ) ≤ Z  f ( x )  p dμ ( x ) ¶ 1 /p Z  g ( x )  q dμ ( x ) ¶ 1 /q . 1 The space L p ( μ ) is a Banach space for p ≥ 1 – this is a normed linear space that is complete – i.e. in which every Cauchy sequence has a limit. The notion of continuity for real valued functions defined on L p ( μ ) is a natural extension of the usual one for Euclidean spaces. A sequence of functions g n converges to a function g in L p ( μ ) if g n g p → 0. A realvalued function ψ defined on L p ( μ ) is said to be continuous if ψ ( g n ) converges to ψ ( g ) as a sequence of real numbers whenever g n converges to g in L p ( μ ). Let’s concretize to a specific example. Let X be the space of positive integers N , A be the power set of N and μ be counting measure that assigns the cardinality of a subset of N as its measure. The space L p ( μ ) is then the space of all real valued sequences { x 1 , x 2 , . . . } that are p ’th power summable – i.e. ∑ ∞ i =1  x i  p < ∞ . Clearly all sequences that have only finitely many nonzero entries satisfy this condition. This space is referred to as the l p space. The l p spaces are infinitedimensional spaces – i.e. these spaces do not have a finite basis....
View
Full
Document
This note was uploaded on 04/14/2010 for the course STATS 612 taught by Professor Moulib during the Winter '08 term at University of Michigan.
 Winter '08
 moulib
 Statistics

Click to edit the document details