The problem of estimating the intensity of a non-stationary Poisson point process arises in many applications. Besides non parametric solutions, e. g. kernel estimators, parametric methods based on maximum likelihood estimation are of interest. In the present paper we have developed an approach in which the parametric function is represented by two-dimensional beta-splines.

In statistics of stochastic processes and random fields, a moment function or a cumulant of an estimate of either the correlation function or the spectral function can often contain an integral involving a cyclic product of kernels. We define and study this class of integrals and prove a Young-Hölder inequality. This inequality further enables us to study asymptotics of the above mentioned integrals in the situation where the kernels depend on a parameter. An application to the problem of estimation...

Let ${𝕖}_t={(e_{t1},\cdots ,e_{tp})}^{\text{'}}$ be a $p$-dimensional nonnegative strict white noise with finite second moments. Let $h_{ij}\left(x\right)$ be nondecreasing functions from $[0,\infty )$ onto $[0,\infty )$ such that $h_{ij}\left(x\right)\le x$ for $i,j=1,\cdots ,p$. Let $𝕌=\left(u_{ij}\right)$ be a $p\times p$ matrix with nonnegative elements having all its roots inside the unit circle. Define a process ${𝕏}_t={(X_{t1},\cdots ,X_{tp})}^{\text{'}}$ by $$X_{tj}=u_{j1}{h}_{1j}\left(X_{t-1,1}\right)+\cdots +u_{jp}{h}_{pj}\left(X_{t-1,p}\right)+e_{tj}$$ for $j=1,\cdots ,p$. A method for estimating $𝕌$ from a realization ${𝕏}_1,\cdots ,{𝕏}_n$ is proposed. It is proved that the estimators are strongly consistent.

Multivariate failure time data arise in various forms including recurrent event data when individuals are followed to observe the sequence of occurrences of a certain type of event; correlated failure time when an individual is followed for the occurrence of two or more types of events for which the individual is simultaneously at risk, or when distinct individuals have depending event times; or more complicated multistate processes where individuals may move among a number of discrete states over...

We study the limit behavior of certain classes of dependent random sequences (processes) which do not possess the Markov property. Assuming these processes depend on a control parameter we show that the optimization of the control can be reduced to a problem of nonlinear optimization. Under certain hypotheses we establish the stability of such optimization problems.