Detecting changes in mean levels with atheoretical regression trees.
In this paper the study of long term changes in lake mean water levels is faced as a problem of detecting multiple structural breaks in the mean occurring at unknown dates. We propose a non parametric approach that exploits, in the framework of least squares regression trees, the contiguity property of the Fisher method proposed for grouping a single real variable. The proposed approach is applied to study the changes in mean water levels of Lake Michigan-Huron.
SubjectsMultiple structural breaks at unknown times
- Engineering: Reports