{"title":"用高斯过程的随机漫步预测用电量","authors":"Chloé Hashimoto-Cullen, Benjamin Guedj","doi":"arxiv-2409.05934","DOIUrl":null,"url":null,"abstract":"We consider time-series forecasting problems where data is scarce, difficult\nto gather, or induces a prohibitive computational cost. As a first attempt, we\nfocus on short-term electricity consumption in France, which is of strategic\nimportance for energy suppliers and public stakeholders. The complexity of this\nproblem and the many levels of geospatial granularity motivate the use of an\nensemble of Gaussian Processes (GPs). Whilst GPs are remarkable predictors,\nthey are computationally expensive to train, which calls for a frugal few-shot\nlearning approach. By taking into account performance on GPs trained on a\ndataset and designing a random walk on these, we mitigate the training cost of\nour entire Bayesian decision-making procedure. We introduce our algorithm\ncalled \\textsc{Domino} (ranDOM walk on gaussIaN prOcesses) and present\nnumerical experiments to support its merits.","PeriodicalId":501425,"journal":{"name":"arXiv - STAT - Methodology","volume":"16 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-09-09","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":"{\"title\":\"Predicting Electricity Consumption with Random Walks on Gaussian Processes\",\"authors\":\"Chloé Hashimoto-Cullen, Benjamin Guedj\",\"doi\":\"arxiv-2409.05934\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"We consider time-series forecasting problems where data is scarce, difficult\\nto gather, or induces a prohibitive computational cost. As a first attempt, we\\nfocus on short-term electricity consumption in France, which is of strategic\\nimportance for energy suppliers and public stakeholders. The complexity of this\\nproblem and the many levels of geospatial granularity motivate the use of an\\nensemble of Gaussian Processes (GPs). Whilst GPs are remarkable predictors,\\nthey are computationally expensive to train, which calls for a frugal few-shot\\nlearning approach. By taking into account performance on GPs trained on a\\ndataset and designing a random walk on these, we mitigate the training cost of\\nour entire Bayesian decision-making procedure. We introduce our algorithm\\ncalled \\\\textsc{Domino} (ranDOM walk on gaussIaN prOcesses) and present\\nnumerical experiments to support its merits.\",\"PeriodicalId\":501425,\"journal\":{\"name\":\"arXiv - STAT - Methodology\",\"volume\":\"16 1\",\"pages\":\"\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2024-09-09\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"0\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"arXiv - STAT - Methodology\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/arxiv-2409.05934\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - STAT - Methodology","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2409.05934","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
摘要
我们考虑的是数据稀缺、难以收集或计算成本过高的时间序列预测问题。作为首次尝试,我们将重点放在法国的短期用电量上,这对能源供应商和公共利益相关者来说具有重要的战略意义。这一问题的复杂性和多级地理空间粒度促使我们使用高斯过程(GPs)组合。虽然 GPs 是出色的预测工具,但其训练的计算成本很高,因此需要一种节俭的少量学习方法。通过考虑在数据集上训练的 GPs 的性能,并在这些 GPs 上设计随机行走,我们减轻了整个贝叶斯决策过程的训练成本。我们介绍了我们的算法,称为textsc{Domino}(ranDOM walk on gaussIaN prOcesses),并通过数值实验来证明它的优点。
Predicting Electricity Consumption with Random Walks on Gaussian Processes
We consider time-series forecasting problems where data is scarce, difficult
to gather, or induces a prohibitive computational cost. As a first attempt, we
focus on short-term electricity consumption in France, which is of strategic
importance for energy suppliers and public stakeholders. The complexity of this
problem and the many levels of geospatial granularity motivate the use of an
ensemble of Gaussian Processes (GPs). Whilst GPs are remarkable predictors,
they are computationally expensive to train, which calls for a frugal few-shot
learning approach. By taking into account performance on GPs trained on a
dataset and designing a random walk on these, we mitigate the training cost of
our entire Bayesian decision-making procedure. We introduce our algorithm
called \textsc{Domino} (ranDOM walk on gaussIaN prOcesses) and present
numerical experiments to support its merits.