Trend detection for heteroscedastic extremes (Q2303026)

The authors examine a model for the extremes of independent but nonstationary observations whose distribution evolves over time, called the model of heteroscedastic extremes. The model setup is the following: Consider a triangular array of independent random variables \(X_{1}^{n},\ldots,X_{n}^{n}\), where \(X_{i}^{n}\) has a continuous distribution function \(F_{n,i} \,, \; n\ge 1\, , \; 1\le i\le n \,\). Assume that all the distributions share the same right endpoint \(x^{*}\) and satisfy the proportional tail condition \[ \lim_{x \to x^{*}} \frac{1-F_{n,i}\left( x\right)}{1-F\left( x\right)} = c\left(\frac{i}{n}\right), \] uniformly in \(n \ge 1 , \; 1 \le i \le n\) , where \(F\) is a baseline continuous distribution function with right endpoint \(x^{*}\) and the skedasis function \(c:[0,1] \to (0,\infty)\) is a continuous function. The authors focus on parametric models for the skedasis function and the corresponding estimation and test problems. From the authors' abstract: ``\dots we \dots prove the consistency and asymptotic normality of the parameter estimators. A parametric test for trend detection in the case where the skedasis function is monotonic is introduced. A short simulation study shows that the parametric test can be more powerful than the nonparametric Kolmogorov-Smirnov type test, even for misspecified models. The methodology is finally illustrated on a dataset of daily maximal temperatures in Fort Collins, Colorado, during the 20th century. ''