{"about":{"site":"https://codewithpapers.app","non_affiliation":"Code with Papers and Syntology are not affiliated with, endorsed by, or sponsored by Papers with Code, Meta, or the pwc-archive mirror.","licence":"CC BY-SA 4.0","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","attribution":"https://codewithpapers.app/attribution","modified":"archive material modified by Syntology; see the attribution page"},"url":"/paper/capturing-structure-implicitly-from-time","title":"Capturing Structure Implicitly from Time-Series having Limited Data","arxiv_id":"1803.05867","date":"2018-03-15","proceeding":null,"authors":["Daniel Emaasit","Matthew Johnson"],"abstract":"Scientific fields such as insider-threat detection and highway-safety\nplanning often lack sufficient amounts of time-series data to estimate\nstatistical models for the purpose of scientific discovery. Moreover, the\navailable limited data are quite noisy. This presents a major challenge when\nestimating time-series models that are robust to overfitting and have\nwell-calibrated uncertainty estimates. Most of the current literature in these\nfields involve visualizing the time-series for noticeable structure and hard\ncoding them into pre-specified parametric functions. This approach is\nassociated with two limitations. First, given that such trends may not be\neasily noticeable in small data, it is difficult to explicitly incorporate\nexpressive structure into the models during formulation. Second, it is\ndifficult to know $\\textit{a priori}$ the most appropriate functional form to\nuse. To address these limitations, a nonparametric Bayesian approach was\nproposed to implicitly capture hidden structure from time series having limited\ndata. The proposed model, a Gaussian process with a spectral mixture kernel,\nprecludes the need to pre-specify a functional form and hard code trends, is\nrobust to overfitting and has well-calibrated uncertainty estimates.","url_abs":"http://arxiv.org/abs/1803.05867v1","url_pdf":"http://arxiv.org/pdf/1803.05867v1.pdf","source":{"archive":"pwc-archive (Hugging Face), CC BY-SA 4.0","snapshot":"2025-07-28","licence_url":"https://creativecommons.org/licenses/by-sa/4.0/legalcode","row_kind":"abstracts"},"code_links":[{"paper_slug":"capturing-structure-implicitly-from-time","repo_url":"https://github.com/emaasit/long-range-extrapolation","is_official":1,"mentioned_in_paper":1,"mentioned_in_github":0,"framework":"none","reach":null}],"tasks":[{"task_slug":"small-data","task_name":"Small Data Image Classification"},{"task_slug":"time-series-1","task_name":"Time Series"},{"task_slug":"time-series","task_name":"Time Series Analysis"},{"task_slug":"scientific-discovery","task_name":"scientific discovery"}],"methods":[{"method_slug":"gaussian-process","method_name":"Gaussian Process"}],"datasets_introduced":[],"methods_introduced":[],"results":[],"syntology":{"atlas_url":null,"mcp":null,"developers":"https://syntology.ai/developers"},"arxiv_metadata":null,"syntology_extracted_results":null}