Properties of Hida processes on \({\mathbb{R}}^ 2\). II. Prediction and interpolation problems for processes on \({\mathbb{R}}^ 2\) (Q2266526)
From MaRDI portal
scientific article
Language | Label | Description | Also known as |
---|---|---|---|
English | Properties of Hida processes on \({\mathbb{R}}^ 2\). II. Prediction and interpolation problems for processes on \({\mathbb{R}}^ 2\) |
scientific article |
Statements
Properties of Hida processes on \({\mathbb{R}}^ 2\). II. Prediction and interpolation problems for processes on \({\mathbb{R}}^ 2\) (English)
0 references
1984
0 references
This paper is a continuation of another article [ibid. 15, 336-360 (1984; Zbl 0552.60047)] where the author has studied the Hida property relative to the family of \(\sigma\)-fields \(\{\) \({\mathcal F}_ z,z\in {\mathbb{R}}^ 2\}\) generated by a two-parameter Wiener process. Now one considers the Hida property with respect to the \(\sigma\)-fields \({\mathcal G}_ z\) corresponding to the values of the Wiener process on the set \(\{\) (u,v):u\(\leq s\) or \(v\leq t\}\), \(z=(s,t)\). The theory is developed for processes allowing a Goursat representation of the form \(Y_ z=\sum_{i}f_ i(z)M_ i(z)\) where \(M_ i(z)\) are martingales. In general, the linear projection of the future of z on \({\mathcal G}_ z\) has infinite dimension and, therefore, the study is restricted to projections of \(p\times q\) lattices. The dimension of these projections is analyzed in relation with the fact that one of the martingales \(M_ i(z)\) is not strong. The author shows that this dimension is bounded by \(N(p+q-1)\) if all martingales are strong. Finally he studies the problem of linear interpolation from the exterior of a rectangle, assuming that \(Y_ z\) is a Gaussian process with a Goursat representation by means of stochastic integrals of kernels that factorize. The author finds an upper bound for the dimension of the projection of a lattice of points inside the rectangle.
0 references
Hida property
0 references
two-parameter Wiener process
0 references
linear interpolation
0 references
stochastic integrals
0 references