2013/12/16 by Stefan Siegert, Jochen Broecker, Siegert, Stefan +4 · 1 citation
Earth and Planetary Sciences · Engineering · Environmental Science · Mathematics · #Applications (stat.AP) #Artificial intelligence #Climate variability and models #Computer science #Data mining #Econometrics #Engineering #Extreme value theory #FOS: Computer and information sciences #Forecast skill #Hydrological Forecasting Using AI #Machine learning #Mathematics #Measure (data warehouse) #Meteorological Phenomena and Simulations #Physics #Precision and recall #Predictive power #Probabilistic logic #Range (aeronautics) #Simple (philosophy) #Statistical model #Statistics #stat.AP
paper · pdf · doi:10.48550/arxiv.1312.4323
published in arXiv (Cornell University), 35-50 (Cornell University) · 32 pages, 13 Figures
arxiv created 2013/12/16 · openalex publication_date 2013/12/16 · arxiv updated 2013/12/17 · openalex created_date 2022/10/01 · openalex updated_date 2026/08/06
We compare probabilistic predictions of extreme temperature anomalies issued\nby two different forecast schemes. One is a dynamical physical weather model,\nthe other a simple data model. We recall the concept of skill scores in order\nto assess the performance of these two different predictors. Although the\nresult confirms the expectation that the (computationally expensive) weather\nmodel outperforms the simple data model, the performance of the latter is\nsurprisingly good. More specifically, for some parameter range, it is even\nbetter than the uncalibrated weather model. Since probabilistic predictions are\nnot easily interpreted by the end user, we convert them into deterministic\nyes/no statements and measure the performance of these by ROC statistics.\nScored in this way, conclusions about model performance partly change, which\nillustrates that predictive power depends on how it is quantified.\n