Event Details

Influence in the world of changepoints

Influence in the world of changepoints

Changepoint models are useful in a variety of contexts to identify points within a time series when something has changed. How these changepoints influence our decision making is important but there is very little practitioners can do to be confident about the existence of changepoints within their data. Within this talk we develop graphical influence diagnostics to characterize the influence of single observations on changepoint inference. We present a framework for investigating instabilities in changepoint segmentations and assessing the influence of single observations on various outputs of a changepoint analysis. We construct graphical diagnostic plots that allow practitioners to assess whether instabilities occur; how and where they occur; and to detect influential individual observations triggering instability. We analyze well-log data to illustrate how such influence diagnostic plots can be used in practice to reveal features of the data that may otherwise remain hidden.