Lineability within probability theory settings (Q2412810): Difference between revisions

The authors consider some classical (counter)examples in probability theory and study the level to which it is possible to obtain lineability-related results. Thus, for instance, the authors recall that there exists a martingale constructed using a Markov chain \(\{X_n : n \in \mathbb N\}\) defined on a probability space \((\Omega,\mathcal F,\mathcal P)\), such that \((X_n(\omega))_n\) converges for every \(\omega \in \Omega\) with \(E[|X_n|]_n \to \infty\) with \(n\) (see, e.g. [\textit{R. B. Ash}, Real analysis and probability. New York, NY: Academic Press (1972; Zbl 1381.28001)]). Here, the authors prove that the set of convergent sequences of random variables \(\{X_n \;:\;n \in \mathbb N\}\) with \(E[|X_n|]_n \to \infty\) is lineable (i.e., the set together with \(0\) contains an infinite-dimensional vector space). Among other results, the following are proved. Theorem. Consider the probability space \((\mathbb R^+, \mathcal B(\mathbb R^+), \lambda)\), where \(\mathcal B\) denotes Borel sets and \(\lambda\) is Lebesgue measure. Then the set of unbounded random variables \(f : \mathbb R \to \mathbb R\) that have a finite expected value is algebrable. (That is, this set, together with \(0\), contains an infinitely generated algebra.) Also: Theorem. Given a probability space \((\Omega,\mathcal F,\mathcal P),\) the set of sequences \(\{X_n : n \in \mathbb N\}\) of mutually independent random variables having zero mean such that \(|\frac{1}{n}\sum_{i=1}^n X_i|\) diverges to \(\infty\) almost surely is lineable.