Steven Y. K. Wong, Jennifer S. K. Chan, Lamiae Azizi
{"title":"Quantifying neural network uncertainty under volatility clustering","authors":"Steven Y. K. Wong, Jennifer S. K. Chan, Lamiae Azizi","doi":"arxiv-2402.14476","DOIUrl":null,"url":null,"abstract":"Time-series with time-varying variance pose a unique challenge to uncertainty\nquantification (UQ) methods. Time-varying variance, such as volatility\nclustering as seen in financial time-series, can lead to large mismatch between\npredicted uncertainty and forecast error. Building on recent advances in neural\nnetwork UQ literature, we extend and simplify Deep Evidential Regression and\nDeep Ensembles into a unified framework to deal with UQ under the presence of\nvolatility clustering. We show that a Scale Mixture Distribution is a simpler\nalternative to the Normal-Inverse-Gamma prior that provides favorable\ncomplexity-accuracy trade-off. To illustrate the performance of our proposed\napproach, we apply it to two sets of financial time-series exhibiting\nvolatility clustering: cryptocurrencies and U.S. equities.","PeriodicalId":501139,"journal":{"name":"arXiv - QuantFin - Statistical Finance","volume":"67 1","pages":""},"PeriodicalIF":0.0000,"publicationDate":"2024-02-22","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"0","resultStr":null,"platform":"Semanticscholar","paperid":null,"PeriodicalName":"arXiv - QuantFin - Statistical Finance","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/arxiv-2402.14476","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
引用次数: 0
Abstract
Time-series with time-varying variance pose a unique challenge to uncertainty
quantification (UQ) methods. Time-varying variance, such as volatility
clustering as seen in financial time-series, can lead to large mismatch between
predicted uncertainty and forecast error. Building on recent advances in neural
network UQ literature, we extend and simplify Deep Evidential Regression and
Deep Ensembles into a unified framework to deal with UQ under the presence of
volatility clustering. We show that a Scale Mixture Distribution is a simpler
alternative to the Normal-Inverse-Gamma prior that provides favorable
complexity-accuracy trade-off. To illustrate the performance of our proposed
approach, we apply it to two sets of financial time-series exhibiting
volatility clustering: cryptocurrencies and U.S. equities.