ORCID Profile
0000-0001-5751-6292
Current Organisations
University of Trento
,
Karolinska University Hospital
,
Fondazione Bruno Kessler
,
Karolinska Institutet
Does something not look right? The information on this page has been harvested from data sources that may not be up to date. We continue to work with information providers to improve coverage and quality. To report an issue, use the Feedback Form.
Publisher: Elsevier BV
Date: 03-2012
Publisher: MDPI AG
Date: 06-12-2019
DOI: 10.3390/RS11242922
Abstract: The use of analog-similar weather patterns for weather forecasting and analysis is an established method in meteorology. The most challenging aspect of using this approach in the context of operational radar applications is to be able to perform a fast and accurate search for similar spatiotemporal precipitation patterns in a large archive of historical records. In this context, sequential pairwise search is too slow and computationally expensive. Here, we propose an architecture to significantly speed up spatiotemporal analog retrieval by combining nonlinear geometric dimensionality reduction (UMAP) with the fastest known Euclidean search algorithm for time series (MASS) to find radar analogs in constant time, independently of the desired temporal length to match and the number of extracted analogs. We show that UMAP, combined with a grid search protocol over relevant hyperparameters, can find analog sequences with lower mean square error (MSE) than principal component analysis (PCA). Moreover, we show that MASS is 20 times faster than brute force search on the UMAP embedding space. We test the architecture on real dataset and show that it enables precise and fast operational analog ensemble search through more than 2 years of radar archive in less than 3 seconds on a single workstation.
Publisher: Elsevier BV
Date: 04-2016
Publisher: Springer Science and Business Media LLC
Date: 13-07-2020
DOI: 10.1038/S41597-020-0574-8
Abstract: We introduce , a high-resolution radar reflectivity dataset collected by the Civil Protection weather radar of the Trentino South Tyrol Region, in the Italian Alps. The dataset includes 894,916 timesteps of precipitation from more than 9 years of data, offering a novel resource to develop and benchmark analog ensemble models and machine learning solutions for precipitation nowcasting. Data are expressed as 2D images, considering the maximum reflectivity on the vertical section at 5 min s ling rate, covering an area of 240 km of diameter at 500 m horizontal resolution. The distribution also includes a curated set of 1,732 sequences, for a total of 362,233 radar images, labeled with precipitation type tags assigned by expert meteorologists. We validate as a benchmark for nowcasting methods by introducing a deep learning model to forecast reflectivity, and a procedure based on the UMAP dimensionality reduction algorithm for interactive exploration. Software methods for data pre-processing, model training and inference, and a pre-trained model are publicly available on GitHub ( github.com/MPBA/TAASRAD19 ) for study replication and reproducibility.
Publisher: Zenodo
Date: 2019
Publisher: Elsevier BV
Date: 10-2011
DOI: 10.1016/J.EURONEURO.2011.08.008
Abstract: The spectrum of disorders of the brain is large, covering hundreds of disorders that are listed in either the mental or neurological disorder chapters of the established international diagnostic classification systems. These disorders have a high prevalence as well as short- and long-term impairments and disabilities. Therefore they are an emotional, financial and social burden to the patients, their families and their social network. In a 2005 landmark study, we estimated for the first time the annual cost of 12 major groups of disorders of the brain in Europe and gave a conservative estimate of €386 billion for the year 2004. This estimate was limited in scope and conservative due to the lack of sufficiently comprehensive epidemiological and/or economic data on several important diagnostic groups. We are now in a position to substantially improve and revise the 2004 estimates. In the present report we cover 19 major groups of disorders, 7 more than previously, of an increased range of age groups and more cost items. We therefore present much improved cost estimates. Our revised estimates also now include the new EU member states, and hence a population of 514 million people. To estimate the number of persons with defined disorders of the brain in Europe in 2010, the total cost per person related to each disease in terms of direct and indirect costs, and an estimate of the total cost per disorder and country. The best available estimates of the prevalence and cost per person for 19 groups of disorders of the brain (covering well over 100 specific disorders) were identified via a systematic review of the published literature. Together with the twelve disorders included in 2004, the following range of mental and neurologic groups of disorders is covered: addictive disorders, affective disorders, anxiety disorders, brain tumor, childhood and adolescent disorders (developmental disorders), dementia, eating disorders, epilepsy, mental retardation, migraine, multiple sclerosis, neuromuscular disorders, Parkinson's disease, personality disorders, psychotic disorders, sleep disorders, somatoform disorders, stroke, and traumatic brain injury. Epidemiologic panels were charged to complete the literature review for each disorder in order to estimate the 12-month prevalence, and health economic panels were charged to estimate best cost-estimates. A cost model was developed to combine the epidemiologic and economic data and estimate the total cost of each disorder in each of 30 European countries (EU27+Iceland, Norway and Switzerland). The cost model was populated with national statistics from Eurostat to adjust all costs to 2010 values, converting all local currencies to Euro, imputing costs for countries where no data were available, and aggregating country estimates to purchasing power parity adjusted estimates for the total cost of disorders of the brain in Europe 2010. The total cost of disorders of the brain was estimated at €798 billion in 2010. Direct costs constitute the majority of costs (37% direct healthcare costs and 23% direct non-medical costs) whereas the remaining 40% were indirect costs associated with patients' production losses. On average, the estimated cost per person with a disorder of the brain in Europe ranged between €285 for headache and €30,000 for neuromuscular disorders. The European per capita cost of disorders of the brain was €1550 on average but varied by country. The cost (in billion €PPP 2010) of the disorders of the brain included in this study was as follows: addiction: €65.7 anxiety disorders: €74.4 brain tumor: €5.2 child/adolescent disorders: €21.3 dementia: €105.2 eating disorders: €0.8 epilepsy: €13.8 headache: €43.5 mental retardation: €43.3 mood disorders: €113.4 multiple sclerosis: €14.6 neuromuscular disorders: €7.7 Parkinson's disease: €13.9 personality disorders: €27.3 psychotic disorders: €93.9 sleep disorders: €35.4 somatoform disorder: €21.2 stroke: €64.1 traumatic brain injury: €33.0. It should be noted that the revised estimate of those disorders included in the previous 2004 report constituted €477 billion, by and large confirming our previous study results after considering the inflation and population increase since 2004. Further, our results were consistent with administrative data on the health care expenditure in Europe, and comparable to previous studies on the cost of specific disorders in Europe. Our estimates were lower than comparable estimates from the US. This study was based on the best currently available data in Europe and our model enabled extrapolation to countries where no data could be found. Still, the scarcity of data is an important source of uncertainty in our estimates and may imply over- or underestimations in some disorders and countries. Even though this review included many disorders, diagnoses, age groups and cost items that were omitted in 2004, there are still remaining disorders that could not be included due to limitations in the available data. We therefore consider our estimate of the total cost of the disorders of the brain in Europe to be conservative. In terms of the health economic burden outlined in this report, disorders of the brain likely constitute the number one economic challenge for European health care, now and in the future. Data presented in this report should be considered by all stakeholder groups, including policy makers, industry and patient advocacy groups, to reconsider the current science, research and public health agenda and define a coordinated plan of action of various levels to address the associated challenges. Political action is required in light of the present high cost of disorders of the brain. Funding of brain research must be increased care for patients with brain disorders as well as teaching at medical schools and other health related educations must be quantitatively and qualitatively improved, including psychological treatments. The current move of the pharmaceutical industry away from brain related indications must be halted and reversed. Continued research into the cost of the many disorders not included in the present study is warranted. It is essential that not only the EU but also the national governments forcefully support these initiatives.
Publisher: Wiley
Date: 12-11-2012
DOI: 10.1016/J.JALZ.2012.06.008
Abstract: The Resource Utilization in Dementia (RUD) questionnaire is the most widely used instrument for resource use data collection in dementia, enabling comparison of costs of care across countries with differing health care provisions. Recent feedback from payers questioned its face validity given that health care provisions have changed since the initial development of the RUD in 1998. The aim of this study was to update the RUD to improve its face validity in Alzheimer's disease (AD) clinical research and its utility for health care resource allocation. An extensive PubMed review was conducted of current relevant resource items in AD in 15 countries. The findings were complemented by interviews with local care providers and experts in dementia care and health economics. Their proposed revisions were discussed with five leading dementia experts in North and South America, northern and southern Europe, and Asia. A new version of the RUD was developed based on their recommendations. RUD users identified a need for more information relevant to coverage decisions. Proposed revisions included changes to existing questions (e.g., to capture more accurately the number and type of health care visits) and the addition of new questions (e.g., on informal caregiver hours and the primary caregiver's hours of sleep). Several minor changes were made to the RUD instrument to improve the accuracy and precision of the data while maintaining comparability with the original version and reflecting current medical practice. The RUD Complete Version 4.0 is now available for use in future AD clinical trials.
Publisher: Wiley
Date: 03-2014
DOI: 10.1111/JOIM.12167
Abstract: The socio-economic impact of Alzheimer's disease (AD) and other dementias is enormous, and the potential economic challenges ahead are clear given the projected future numbers of in iduals with these conditions. Because of the high prevalence and cost of dementia, it is very important to assess any intervention from a cost-effectiveness viewpoint. The diagnostic criteria for preclinical AD suggested by the National Institute on Aging and Alzheimer's Association workgroups in combination with the goal of effective disease-modifying treatment (DMT) are, however, a challenge for clinical practice and for the design of clinical trials. Key issues for future cost-effectiveness studies include the following: (i) the consequences for patients if diagnosis is shifted from AD-dementia to predementia states, (ii) bridging the gap between clinical trial populations and patients treated in clinical practice, (iii) translation of clinical trial end-points into measures that are meaningful to patients and policymakers ayers and (iv) how to measure long-term effects. To improve cost-effectiveness studies, long-term population-based data on disease progression, costs and outcomes in clinical practice are needed not only in dementia but also in predementia states. Reliable surrogate end-points in clinical trials that are sensitive to detect effects even in predementia states are also essential as well as robust and validated modelling methods from predementia states that also take into account comorbidities and age. Finally, the ethical consequences of early diagnosis should be considered.
Publisher: Wiley
Date: 19-12-2011
DOI: 10.1111/J.1468-1331.2011.03590.X
Abstract: In 2005, we presented for the first time overall estimates of annual costs for brain disorders (mental and neurologic disorders) in Europe. This new report presents updated, more accurate, and comprehensive 2010 estimates for 30 European countries. One-year prevalence and annual cost per person of 19 major groups of disorders are based on 'best estimates' derived from systematic literature reviews by panels of experts in epidemiology and health economics. Our cost estimation model was populated with national statistics from Eurostat to adjust to 2010 values, converting all local currencies to Euros (€), imputing cost for countries where no data were available, and aggregating country estimates to purchasing power parity-adjusted estimates of the total cost of brain disorders in Europe in 2010. Total European 2010 cost of brain disorders was €798 billion, of which direct health care cost 37%, direct non-medical cost 23%, and indirect cost 40%. Average cost per inhabitant was €5.550. The European average cost per person with a disorder of the brain ranged between €285 for headache and €30 000 for neuromuscular disorders. Total annual cost per disorder (in billion € 2010) was as follows: addiction 65.7 anxiety disorders 74.4 brain tumor 5.2 child/adolescent disorders 21.3 dementia 105.2 eating disorders 0.8 epilepsy 13.8 headache 43.5 mental retardation 43.3 mood disorders 113.4 multiple sclerosis 14.6 neuromuscular disorders 7.7 Parkinson's disease 13.9 personality disorders 27.3 psychotic disorders 93.9 sleep disorders 35.4 somatoform disorder 21.2 stroke 64.1 and traumatic brain injury 33.0. Our cost model revealed that brain disorders overall are much more costly than previously estimated constituting a major health economic challenge for Europe. Our estimate should be regarded as conservative because many disorders or cost items could not be included because of lack of data.
Publisher: MDPI AG
Date: 07-03-2020
Abstract: One of the most crucial applications of radar-based precipitation nowcasting systems is the short-term forecast of extreme rainfall events such as flash floods and severe thunderstorms. While deep learning nowcasting models have recently shown to provide better overall skill than traditional echo extrapolation models, they suffer from conditional bias, sometimes reporting lower skill on extreme rain rates compared to Lagrangian persistence, due to excessive prediction smoothing. This work presents a novel method to improve deep learning prediction skills in particular for extreme rainfall regimes. The solution is based on model stacking, where a convolutional neural network is trained to combine an ensemble of deep learning models with orographic features, doubling the prediction skills with respect to the ensemble members and their average on extreme rain rates, and outperforming them on all rain regimes. The proposed architecture was applied on the recently released TAASRAD19 radar dataset: the initial ensemble was built by training four models with the same TrajGRU architecture over different rainfall thresholds on the first six years of the dataset, while the following three years of data were used for the stacked model. The stacked model can reach the same skill of Lagrangian persistence on extreme rain rates while retaining superior performance on lower rain regimes.
No related grants have been discovered for Linus Jönsson.