On maxima of partial samples in Gaussian sequences with pseudo-stationary trends (Q2471660): Difference between revisions

Let \(X_n\), \(n\in\mathbb{Z}\), be a Gaussian stationary centered sequence with unit variance, and let \(m_n\), \(n\in\mathbb{Z}\), be a nonrandom sequence. This paper studies the joint distribution of the two random variables \[ \max_{i=1,\dots,n}(X_i+m_id_n),\quad \max_{i\in G_n}(X_i+m_id_n), \] as \(n\) increases, where \(G_n\) is a subset of \(\{1,\dots,n\}\). The sequence \(d_n\) tends to zero in such a way that under appropriate normalization, the limit of the joint distribution will exist. This study is motivated by problems of statistics of extremes for correlated and non identically distributed samples when some data are missing.