Jonas Wallin
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Jonas Wallin has a total of 57 co-authors. Here are the top co-authors by number of shared publications:
- Alexandre B. Simas
Statistics Program, King Abdullah University of Science and Technology (KAUST) , Thuwal 23955
5 shared publications - David Bolin
Statistics Program, King Abdullah University of Science and Technology (KAUST) , Thuwal 23955
5 shared publications David Bolin
King Abdullah University of Science and Technology , Thuwal , Saudi Arabia
3 shared publications- Malgorzata Bogdan
University of Wroclaw Wroclaw Poland
3 shared publications - Joacim Rocklöv
Heidelberg University
2 shared publications - Johan Larsson
Dept. of Statistics Dept. of Mathematics Dept. of Statistics Dept. of Statistics Lund University University of Wroclaw Lund University Lund University
2 shared publications
Latest Publications
- 6 July 2026
- Published by American Psychological Association (APA) in Psychological Methods
Heterogeneous variance models with Gaussian processes.
Understanding variability is a key focus in many areas of psychological research, with growing interest in modeling individual- and group-level variability. Although multilevel models such as heterogeneous variance models (HVMs) and Mixed-Effects Location-Scale models have been used to capture these dynamics, they typically rely on linear assumptions and restrict temporal changes in variability to a single level. Recent calls for nonlinear approaches in psychology highlight the need for more flexible models that can better account for complex, dynamic processes. This article introduces the use of Gaussian processes (GPs) within the framework of HVMs to address these limitations. By incorporating GPs in HVMs, we allow for the modeling of nonlinear variability across multiple levels, including temporal dynamics at both the individual and group levels. We demonstrate the benefits of this approach in two empirical applications. Our findings show that using GPs provides an improved model fit compared with traditional linear methods and highlight the utility of GPs in variance modeling, offering new possibilities for studying dynamic and emergent processes in psychological and social science research. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
- 6 July 2026
- Published by American Psychological Association (APA) in Psychological Methods
Heterogeneous variance models with Gaussian processes.
Understanding variability is a key focus in many areas of psychological research, with growing interest in modeling individual- and group-level variability. Although multilevel models such as heterogeneous variance models (HVMs) and Mixed-Effects Location-Scale models have been used to capture these dynamics, they typically rely on linear assumptions and restrict temporal changes in variability to a single level. Recent calls for nonlinear approaches in psychology highlight the need for more flexible models that can better account for complex, dynamic processes. This article introduces the use of Gaussian processes (GPs) within the framework of HVMs to address these limitations. By incorporating GPs in HVMs, we allow for the modeling of nonlinear variability across multiple levels, including temporal dynamics at both the individual and group levels. We demonstrate the benefits of this approach in two empirical applications. Our findings show that using GPs provides an improved model fit compared with traditional linear methods and highlight the utility of GPs in variance modeling, offering new possibilities for studying dynamic and emergent processes in psychological and social science research. (PsycInfo Database Record (c) 2026 APA, all rights reserved).
- Open Access
- 11 May 2026
- Published by Oxford University Press (OUP) in Journal of the Royal Statistical Society Series B: Statistical Methodology
- Vol. 88 (5) , 1805-1823
Statistical inference for Gaussian Whittle–Matérn fields on metric graphs
Whittle–Matérn fields are a recently introduced class of Gaussian processes on metric graphs, specified as solutions to a fractional-order stochastic differential equation. Unlike previous covariance-based methods, these fields are well-defined for any compact metric graph and can provide Gaussian processes with differentiable sample paths. We derive the main statistical properties, including the consistency and asymptotic normality of maximum likelihood estimators and the necessary and sufficient conditions for optimal prediction with misspecified parameters. The covariance function is generally unavailable in closed form, which makes statistical inference challenging. However, we show that for specific values of the fractional exponent where the fields exhibit Markov properties, likelihood-based inference and spatial prediction can be performed exactly and efficiently. This enables the use of Whittle–Matérn fields in large datasets without approximations. The results and methods are illustrated through simulation studies and through an application to traffic data modelling, where allowing for differentiable processes significantly improves results.
- 11 May 2026
- Published by Oxford University Press (OUP) in Journal of the Royal Statistical Society Series B: Statistical Methodology
- Vol. 88 (5) , 1805-1823
Open AccessStatistical inference for Gaussian Whittle–Matérn fields on metric graphs
Whittle–Matérn fields are a recently introduced class of Gaussian processes on metric graphs, specified as solutions to a fractional-order stochastic differential equation. Unlike previous covariance-based methods, these fields are well-defined for any compact metric graph and can provide Gaussian processes with differentiable sample paths. We derive the main statistical properties, including the consistency and asymptotic normality of maximum likelihood estimators and the necessary and sufficient conditions for optimal prediction with misspecified parameters. The covariance function is generally unavailable in closed form, which makes statistical inference challenging. However, we show that for specific values of the fractional exponent where the fields exhibit Markov properties, likelihood-based inference and spatial prediction can be performed exactly and efficiently. This enables the use of Whittle–Matérn fields in large datasets without approximations. The results and methods are illustrated through simulation studies and through an application to traffic data modelling, where allowing for differentiable processes significantly improves results.
- 5 March 2026
- Published by Elsevier in Stochastic Processes and their Applications
- Vol. 196
An explicit link between graphical models and Gaussian Markov random fields on metric graphs
- 5 March 2026
- Published by Elsevier in Stochastic Processes and their Applications
- Vol. 196
An explicit link between graphical models and Gaussian Markov random fields on metric graphs
- Open Access
Scalable Ultra-High-Dimensional Quantile Regression with Genomic Applications
Modern datasets arising from social media, genomics, and biomedical informatics are often heterogeneous and (ultra) high-dimensional, creating substantial challenges for conventional modeling techniques. Quantile regression (QR) not only offers a flexible way to capture heterogeneous effects across the conditional distribution of an outcome, but also naturally produces prediction intervals that help quantify uncertainty in future predictions. However, classical QR methods can face serious memory and computational constraints in large-scale settings. These limitations motivate the use of parallel computing to maintain tractability. While extensive work has examined sample-splitting strategies in settings where the number of observations $n$ greatly exceeds the number of features $p$, the equally important (ultra) high-dimensional regime ($p >> n$) has been comparatively underexplored. To address this gap, we introduce a feature-splitting proximal point algorithm, FS-QRPPA, for penalized QR in high-dimensional regime. Leveraging recent developments in variational analysis, we establish a Q-linear convergence rate for FS-QRPPA and demonstrate its superior scalability in large-scale genomic applications from the UK Biobank relative to existing methods. Moreover, FS-QRPPA yields more accurate coefficient estimates and better coverage for prediction intervals than current approaches. We provide a parallel implementation in the R package fsQRPPA, making penalized QR tractable on large-scale datasets.Open AccessScalable Ultra-High-Dimensional Quantile Regression with Genomic Applications
Modern datasets arising from social media, genomics, and biomedical informatics are often heterogeneous and (ultra) high-dimensional, creating substantial challenges for conventional modeling techniques. Quantile regression (QR) not only offers a flexible way to capture heterogeneous effects across the conditional distribution of an outcome, but also naturally produces prediction intervals that help quantify uncertainty in future predictions. However, classical QR methods can face serious memory and computational constraints in large-scale settings. These limitations motivate the use of parallel computing to maintain tractability. While extensive work has examined sample-splitting strategies in settings where the number of observations $n$ greatly exceeds the number of features $p$, the equally important (ultra) high-dimensional regime ($p >> n$) has been comparatively underexplored. To address this gap, we introduce a feature-splitting proximal point algorithm, FS-QRPPA, for penalized QR in high-dimensional regime. Leveraging recent developments in variational analysis, we establish a Q-linear convergence rate for FS-QRPPA and demonstrate its superior scalability in large-scale genomic applications from the UK Biobank relative to existing methods. Moreover, FS-QRPPA yields more accurate coefficient estimates and better coverage for prediction intervals than current approaches. We provide a parallel implementation in the R package fsQRPPA, making penalized QR tractable on large-scale datasets. - Open Access
Modelling serological multiplex bead assays responses: A case study from Malaysia
Background: Multiplex bead assays (MBAs) provide quantitative measurements of many analytes from small sample volumes, reducing cost and processing time compared with traditional immunoassays. These advantages have made MBAs valuable for studying diverse diseases, particularly in low-resource settings. However, most analytical approaches focus on individual diseases, while integrated surveillance platforms would benefit from methods that jointly analyse the full range of pathogens included in multiplex assays. Methods: We applied factor analysis combined with a Dirichlet process mixture model to identify sub-populations based on MBA responses and assess whether these groups show spatial patterns or share socioeconomic characteristics and disease exposures. Data were drawn from four districts in northern Sabah, Malaysia, and included antibody responses for several neglected tropical diseases (NTDs): strongyloidiasis, lymphatic filariasis, giardiasis, toxoplasmosis, trachoma, and yaws. Results: The model identified four distinct sub-populations. Three of these were spatially distributed and included people with similar socioeconomic profiles. These shared characteristics may help explain the antibody patterns observed within each group, offering a comprehensive characterization of each sub-population. Conclusion: This methodology replicable to other MBA datasets and can provide new insights into the interplay of various exposures at the population level. It also has potential to strengthen integrated surveillance systems by informing more targeted sampling strategies.
Open AccessModelling serological multiplex bead assays responses: A case study from Malaysia
Background: Multiplex bead assays (MBAs) provide quantitative measurements of many analytes from small sample volumes, reducing cost and processing time compared with traditional immunoassays. These advantages have made MBAs valuable for studying diverse diseases, particularly in low-resource settings. However, most analytical approaches focus on individual diseases, while integrated surveillance platforms would benefit from methods that jointly analyse the full range of pathogens included in multiplex assays. Methods: We applied factor analysis combined with a Dirichlet process mixture model to identify sub-populations based on MBA responses and assess whether these groups show spatial patterns or share socioeconomic characteristics and disease exposures. Data were drawn from four districts in northern Sabah, Malaysia, and included antibody responses for several neglected tropical diseases (NTDs): strongyloidiasis, lymphatic filariasis, giardiasis, toxoplasmosis, trachoma, and yaws. Results: The model identified four distinct sub-populations. Three of these were spatially distributed and included people with similar socioeconomic profiles. These shared characteristics may help explain the antibody patterns observed within each group, offering a comprehensive characterization of each sub-population. Conclusion: This methodology replicable to other MBA datasets and can provide new insights into the interplay of various exposures at the population level. It also has potential to strengthen integrated surveillance systems by informing more targeted sampling strategies.
- 4 November 2025
- Published by Springer Nature in Annals of the Institute of Statistical Mathematics
- Vol. 78 (3) , 479-514
Unveiling low-dimensional patterns induced by convex non-differentiable regularizers
- 4 November 2025
- Published by Springer Nature in Annals of the Institute of Statistical Mathematics
- Vol. 78 (3) , 479-514
Unveiling low-dimensional patterns induced by convex non-differentiable regularizers
- Open Access
An explicit link between graphical models and Gaussian Markov random fields on metric graphs
We derive an explicit link between Gaussian Markov random fields on metric graphs and graphical models, and in particular show that a Markov random field restricted to the vertices of the graph is, under mild regularity conditions, a Gaussian graphical model with a distribution which is faithful to its pairwise independence graph, which coincides with the neighbor structure of the metric graph. This is used to show that there are no Gaussian random fields on general metric graphs which are both Markov and isotropic in some suitably regular metric on the graph, such as the geodesic or resistance metrics.
Open AccessAn explicit link between graphical models and Gaussian Markov random fields on metric graphs
We derive an explicit link between Gaussian Markov random fields on metric graphs and graphical models, and in particular show that a Markov random field restricted to the vertices of the graph is, under mild regularity conditions, a Gaussian graphical model with a distribution which is faithful to its pairwise independence graph, which coincides with the neighbor structure of the metric graph. This is used to show that there are no Gaussian random fields on general metric graphs which are both Markov and isotropic in some suitably regular metric on the graph, such as the geodesic or resistance metrics.
- Open Access
Disclosing temperature sensitivity of West Nile virus transmission: novel computational approaches to mosquito-pathogen trait responses
Temperature influences the transmission of mosquito-borne pathogens with significant implications for disease risk in the context of climate change. Mathematical models of mosquito-borne infections rely on functions that capture mosquito-pathogen interactions in response to temperature to better estimate transmission dynamics. To derive these functions, experimental studies provide valuable data on the temperature sensitivity of mosquito life-history traits and pathogen transmission. However, the scarcity of experimental data and inconsistencies in methodologies for analysing and comparing temperature responses across mosquito species and pathogens present challenges to accurately modelling mosquito-borne infections. In this study, we investigated the thermal biology of West Nile virus (WNV), a major mosquito-borne infection. We critically reviewed existing experimental studies, obtaining temperature responses for eight mosquito-pathogen traits across 15 mosquito species. Using this data, we employed Bayesian hierarchical models to estimate temperature response functions for each trait and estimate their variation between species and experiments. We incorporate the resulting functions into mathematical models to estimate the temperature sensitivity of WNV transmission, focusing on six competent mosquito species of the genus Culex . Our study finds similarities in the temperature response among Culex species, with a general optimal transmission temperature around 24°C. We demonstrate that differing mechanistic assumptions in published models can result in a temperature optimum variation exceeding 3°C, underscoring the need for model scrutiny. Additionally, we identify substantial variability between trait temperature responses across experiments on the same mosquito species, indicating significant intra-species variation in trait performance. We provide recommendations for future experimental studies, emphasizing the critical need for additional data on all mosquito-pathogen traits, except for mosquito larva-to-adult development rate and survival, which are relatively well-documented. Incorporating additional data into our multi-species approach enhances the accuracy of temperature trait response estimates. These improvements are critical for forecasting future shifts in mosquito-borne disease risk due to climate change.
Open AccessDisclosing temperature sensitivity of West Nile virus transmission: novel computational approaches to mosquito-pathogen trait responses
Temperature influences the transmission of mosquito-borne pathogens with significant implications for disease risk in the context of climate change. Mathematical models of mosquito-borne infections rely on functions that capture mosquito-pathogen interactions in response to temperature to better estimate transmission dynamics. To derive these functions, experimental studies provide valuable data on the temperature sensitivity of mosquito life-history traits and pathogen transmission. However, the scarcity of experimental data and inconsistencies in methodologies for analysing and comparing temperature responses across mosquito species and pathogens present challenges to accurately modelling mosquito-borne infections. In this study, we investigated the thermal biology of West Nile virus (WNV), a major mosquito-borne infection. We critically reviewed existing experimental studies, obtaining temperature responses for eight mosquito-pathogen traits across 15 mosquito species. Using this data, we employed Bayesian hierarchical models to estimate temperature response functions for each trait and estimate their variation between species and experiments. We incorporate the resulting functions into mathematical models to estimate the temperature sensitivity of WNV transmission, focusing on six competent mosquito species of the genus Culex . Our study finds similarities in the temperature response among Culex species, with a general optimal transmission temperature around 24°C. We demonstrate that differing mechanistic assumptions in published models can result in a temperature optimum variation exceeding 3°C, underscoring the need for model scrutiny. Additionally, we identify substantial variability between trait temperature responses across experiments on the same mosquito species, indicating significant intra-species variation in trait performance. We provide recommendations for future experimental studies, emphasizing the critical need for additional data on all mosquito-pathogen traits, except for mosquito larva-to-adult development rate and survival, which are relatively well-documented. Incorporating additional data into our multi-species approach enhances the accuracy of temperature trait response estimates. These improvements are critical for forecasting future shifts in mosquito-borne disease risk due to climate change.
- Open Access
Statistical inference for Gaussian Whittle-Matérn fields on metric graphs
Whittle-Matérn fields are a recently introduced class of Gaussian processes on metric graphs, which are specified as solutions to a fractional-order stochastic differential equation. Unlike earlier covariance-based approaches for specifying Gaussian fields on metric graphs, the Whittle-Matérn fields are well-defined for any compact metric graph and can provide Gaussian processes with differentiable sample paths. We derive the main statistical properties of the model class, particularly the consistency and asymptotic normality of maximum likelihood estimators of model parameters and the necessary and sufficient conditions for asymptotic optimality properties of linear prediction based on the model with misspecified parameters. The covariance function of the Whittle-Matérn fields is generally unavailable in closed form, and they have therefore been challenging to use for statistical inference. However, we show that for specific values of the fractional exponent, when the fields have Markov properties, likelihood-based inference and spatial prediction can be performed exactly and computationally efficiently. This facilitates using the Whittle-Matérn fields in statistical applications involving big datasets without the need for any approximations. The methods are illustrated via an application to modeling of traffic data, where allowing for differentiable processes dramatically improves the results.
Open AccessStatistical inference for Gaussian Whittle-Matérn fields on metric graphs
Whittle-Matérn fields are a recently introduced class of Gaussian processes on metric graphs, which are specified as solutions to a fractional-order stochastic differential equation. Unlike earlier covariance-based approaches for specifying Gaussian fields on metric graphs, the Whittle-Matérn fields are well-defined for any compact metric graph and can provide Gaussian processes with differentiable sample paths. We derive the main statistical properties of the model class, particularly the consistency and asymptotic normality of maximum likelihood estimators of model parameters and the necessary and sufficient conditions for asymptotic optimality properties of linear prediction based on the model with misspecified parameters. The covariance function of the Whittle-Matérn fields is generally unavailable in closed form, and they have therefore been challenging to use for statistical inference. However, we show that for specific values of the fractional exponent, when the fields have Markov properties, likelihood-based inference and spatial prediction can be performed exactly and computationally efficiently. This facilitates using the Whittle-Matérn fields in statistical applications involving big datasets without the need for any approximations. The methods are illustrated via an application to modeling of traffic data, where allowing for differentiable processes dramatically improves the results.
- Open Access
Gaussian Whittle-Matérn fields on metric graphs
We define a new class of Gaussian processes on compact metric graphs such as street or river networks. The proposed models, the Whittle--Mat\'ern fields, are defined via a fractional stochastic differential equation on the compact metric graph and are a natural extension of Gaussian fields with Mat\'ern covariance functions on Euclidean domains to the non-Euclidean metric graph setting. Existence of the processes, as well as some of their main properties, such as sample path regularity are derived. The model class in particular contains differentiable processes. To the best of our knowledge, this is the first construction of a differentiable Gaussian process on general compact metric graphs. Further, we prove an intrinsic property of these processes: that they do not change upon addition or removal of vertices with degree two. Finally, we obtain Karhunen--Lo\`eve expansions of the processes, provide numerical experiments, and compare them to Gaussian processes with isotropic covariance functions.
Open AccessGaussian Whittle-Matérn fields on metric graphs
We define a new class of Gaussian processes on compact metric graphs such as street or river networks. The proposed models, the Whittle--Mat\'ern fields, are defined via a fractional stochastic differential equation on the compact metric graph and are a natural extension of Gaussian fields with Mat\'ern covariance functions on Euclidean domains to the non-Euclidean metric graph setting. Existence of the processes, as well as some of their main properties, such as sample path regularity are derived. The model class in particular contains differentiable processes. To the best of our knowledge, this is the first construction of a differentiable Gaussian process on general compact metric graphs. Further, we prove an intrinsic property of these processes: that they do not change upon addition or removal of vertices with degree two. Finally, we obtain Karhunen--Lo\`eve expansions of the processes, provide numerical experiments, and compare them to Gaussian processes with isotropic covariance functions.