Even though most work on change-point estimation focuses on changes in the mean, changes in the variance or in the tail distribution can lead to more extreme events. In this paper, we develop a new method of detecting and estimating the change-points in the tail of multiple time series data. In addition, we adapt existing tail change-point detection methods to our specific problem and conduct a...

When testing hypotheses, rare or unexpected observations are normatively more informative than common observations, and recent studies have shown that participants’ behavior reflects this principle. Research has also shown that, when asked to test conditional hypotheses (“If X, then Y”) that are abstract or unfamiliar, participants overwhelmingly consider a supporting observation mentioned in t...

The collapse of the roof of the Baltimore & Ohio (B&O) Railroad Museum Roundhouse onto its collections during a snowstorm in 2003 provides a starting point for our exploration of the link between learning and rare events. The collapse occurred as the museum was preparing for another rare event: the Fair of the Iron Horse, an event planned to celebrate the 175th anniversary of American railroadi...

(b) In our original analysis of spontaneous fronts (section 2.2), we required that ∆(t) = O( ), and Taylor expanded with respect to ∆(t). The correct condition is d∆(t) = O( ). Since ∆(t) turns out to be a Wiener process, this means that d∆(t) is distributed according to a Gaussian and could take on arbitrarily large values. However, for weak noise the probability that this occurs is exponentia...

At very high energy the same universal relation between the multiparticle or the transverse energy distribution associated to a rare event C, P C and the corresponding minimum bias distribution P, P C (ν) ≡ ν/ < ν > P (ν), ν ≡ n or E T works for nucleus-nucleus collisions as well as for hadron-hadron collisions. This suggests that asymptotically, all hadronic processes are similar.

Max-stable processes (de Haan, 1984) have received sustained attention in recent years because of their relevance for studying extreme events in financial, environmental and climate sciences. In a seminal unpublished University of Surrey 1990 technical report, R. L. Smith defined Gaussian max-stable processes, where all margins follow a unit Fréchet distribution, in view of modelling spatial ex...

Four scintillation paddles and coincidence techniques are used to determine the overall flux and angular distribution of cosmic ray muons. The muon lifetime is measured using rare events where, after passage of a muon into a scintillator is detected, its decay is also detected a short time later. The distribution of the decay times provides information about the average muon lifetime. Statistic...

Statistical models for the occurrence of extreme or rare events has been applied in a variety of areas, but little work has been done in extending these ideas to spatial data sets. Here, a brief introduction to the theory of extreme-value statistics is given, and a hierarchical Bayesian framework is used to incorporate a spatial component into the model. Two examples are used to motivate the ap...

Building on earlier work that integrates different factors in language modeling, we view (i) backing off to a shorter history and (ii) class-based generalization as two complementary mechanisms of using a larger equivalence class for prediction when the default equivalence class is too small for reliable estimation. This view entails that the classes in a language model should be learned from r...