arXiv subjects
Publications and source records attributed to Elena Shmileva.
This work is devoted to the clustering of check-in sequences from a geosocial network. We used the mixture Markov chain process as a mathematical model for time-dependent types of data. For clustering, we adjusted the Expectation-Maximization (EM) algorithm. As a result, we obtained highly detailed communities (clusters) of users of the now defunct geosocial network, Weeplaces.
We introduce basic statistical methods for the extrapolation of stationary random fields. For square integrable fields, we set out basics of the kriging extrapolation techniques. For (non--Gaussian) stable fields, which are known to be heavy tailed, we describe further extrapolation methods and discuss their properties. Two of them can be seen as direct generalizations of kriging.
In this paper, we discuss three extrapolation methods for alpha-stable random fields with 1<alpha<=2. We justify them, giving proofs of the existence and uniqueness of the solutions for each method and providing sufficient conditions for path continuity. Two methods are based on minimizing the variability of the difference between the predictor and the theoretical value, whereas in the third approach we provide a new method that maximizes the covariation between these two quantities.
Consider a symmetric $α$-stable Lévy process with $α\in (1,2)$. We study shifted small ball probabilities for these processes in the uniform topology, when the shift function is an arbitrary continuous function which starts at 0. We obtain the exact rate of decrease for these probabilities including constants. Using these small ball estimates, we obtain a functional LIL for $α$-stable Lévy process with attracting functions that are continuous. It occurs that the limit set for the family of renormalized $α$-stable Lévy processes is equal to the set of all continuous functions on $[0,1]$ which start at 0, under certain choice of normalizing functions.