\(\alpha\)-Chaos (Q1116551)

An \(\alpha\)-chaos is a random process with second-order moment structure that of Brownian motion but satisfying a condition that is weaker than that of independent increments. The condition is that the increments of the process form sub-\(\gamma\)-systems for all \(\gamma >\alpha\) and that \(\alpha\) is maximal for this condition. Here the notion of a sub- \(\gamma\)-system is a technical one related to that of \(\Lambda\) (p)-sets in harmonic analysis; it measures lack of independence. Brownian motion provides an example of a 1-chaos. Using stochastic integration techniques it is shown that \(\alpha\)-chaos processes have sample paths which are almost surely continuous and of unbounded variation. The existence of \(\alpha\)-chaoses is established for all \(\alpha >1\).