2018/12/23 by Forrest Paton, Paul D. McNicholas, Paton, Forrest +1 · 1 citation
Computer Science · Mathematics · Physics and Astronomy · #Applications (stat.AP) #Data Analysis with R #FOS: Computer and information sciences #Machine Learning (stat.ML) #Methodology (stat.ME) #Scientific Research and Discoveries #Statistical Methods and Applications
paper · pdf · doi:10.48550/arxiv.1812.09758
openalex publication_date 2018/12/23 · openalex created_date 2022/08/01 · openalex updated_date 2026/07/28
Functional data analysis is a statistical framework where data are assumed to\nfollow some functional form. This method of analysis is commonly applied to\ntime series data, where time, measured continuously or in discrete intervals,\nserves as the location for a function's value. Gaussian processes are a\ngeneralization of the multivariate normal distribution to function space and,\nin this paper, they are used to shed light on coastal rainfall patterns in\nBritish Columbia (BC). Specifically, this work addressed the question over how\none should carry out an exploratory cluster analysis for the BC, or any\nsimilar, coastal rainfall data. An approach is developed for clustering\nmultiple processes observed on a comparable interval, based on how similar\ntheir underlying covariance kernel is. This approach provides interesting\ninsights into the BC data, and these insights can be framed in terms of El\nNi ~no and La Ni ~na; however, the result is not simply one cluster\nrepresenting El Ni ~no years and another for La Ni ~na years. From one\nperspective, the results show that clustering annual rainfall can potentially\nbe used to identify extreme weather patterns.\n