Weak parallelogram laws on Banach spaces and applications to prediction (Q313461)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Weak parallelogram laws on Banach spaces and applications to prediction |
scientific article |
Statements
Weak parallelogram laws on Banach spaces and applications to prediction (English)
0 references
9 September 2016
0 references
The weak parallelogram laws studied in this paper take the form \[ \| x+y\| ^p + C \| x-y\| ^p \leq 2^{p-1} (\| x\| ^p+\| y\| ^p) \] or \[ \| x+y\| ^p + C \| x-y\| ^p \geq 2^{p-1} (\| x\| ^p+\| y\| ^p) \] for all elements of a Banach space \(X\). In the first case \(X\) is said to satisfy a lower weak parallelogram law with constants \(C>0\) and \(p>1\) (\(p\)-LWP(\(C\)) for short), and in the second case \(X\) is said to satisfy an upper weak parallelogram law with constants \(C>0\) and \(p>1\) (\(p\)-UWP(\(C\)) for short). The Clarkson inequalities are of this type; precisely, for \(1<p\leq2\), \(L_p\) satisfies \(p\)-UWP(\(1\)) and for \(2\leq p<\infty\), \(L_p\) satisfies \(p\)-LWP(\(1\)). Another example are the Bynum inequalities. The authors exhibit for precisely which \(r\) the space \(L_p\) satisfies \(r\)-LWP(\(C\)), resp.\ \(r\)-UWP(\(C\)), for some \(C\). They go on to study the geometry of spaces satisfying an LWP or UWP, and they rephrase the defining inequalities for vectors that are orthogonal in the sense of Birkhoff and James. Further, they study orthogonal series (in this sense) in such spaces. For such series an inequality of Baxter type is derived, which has some applications in the prediction theory of stochastic processes.
0 references
parallelogram law
0 references
Pythagorean theorem
0 references
uniform convexity
0 references
best predictor
0 references
Baxter's inequality
0 references
purely nondeterministic
0 references
0 references
0 references