Multi-region Models Research Articles

Multi-region models built with machine and deep learning for predicting several heat-related health outcomes

As a result of climate change, populations worldwide will be exposed to more heat episodes. To ensure a sustainable future, cutting-edge tools must be developed to predict the health effects of heat and limit its consequences. However, current research has mainly focused on one health outcome in a single city/region, thus providing limited knowledge to improve society's resilience to extreme heat. In this study, a machine learning (ML) framework is introduced to predict several heat-related health outcomes in multiple regions simultaneously, using the province of Quebec (Canada) as a case study. Five ML models including penalized regression, ensemble tree-based models and deep neural networks were considered and compared. Models were trained to predict these health outcomes using various meteorological, regional and temporal predictors across all regions. Our results showed that deep learning models were the most promising, with out-of-sample R2 of >60 % for most of the studied health outcomes. However, ensemble tree-based approaches also had the best performance for some health outcomes, and were more sensitive to weather variables and to heatwaves. By introducing novel ML-based tools for predicting heat risks in several regions, this study can guide climate change adaptation and help cities and society to become more healthy, resilient and sustainable.

Abstract 3096: Multi-region patient-derived xenograft models from non-small cell lung cancer patients enrolled in lung TRACERx

Abstract Patient-derived xenograft (PDX) models have become important models in cancer biology. They are thought to mimic tumor biology more closely than traditional cell lines as a consequence of their in vivo heterocellularity and cell-matrix interactions, their 3D architecture and their relatively recent derivation. The genomic fidelity of PDXs is integral to their applications in both basic research and in translational medicine. We derived multi-region PDX models from patients enrolled in TRACERx - a study of the evolutionary dynamics of non-small cell lung cancer (NSCLC) that uses a multi-region deep whole-exome sequencing (WES) approach - using more extensive spatial sampling to better understand the histological and genomic fidelity of the PDX approach. We transplanted regional NSCLC tumor tissue subcutaneously into immunocompromised NSG mice. 134 regions were attempted, resulting in 60 passagable xenografts (44.8%). Of these, 44 regional xenografts (32.8% of all regions) were NSCLC-derived, while 16 (11.9% of all regions) were B-cell lymphoproliferative disease. Multi-region sampling revealed heterogeneous success of PDX derivation between regions within individual patient tumors. There was correlation between PDX derivation and tumor purity by WES. Overall, NSCLC PDX models were established for 22 patients (50%; range 1-5 regional models per patient; Table 1). Histologically, broad similarity was observed between PDX models and corresponding patient tumor regions, although in a small number of models, variability was noted between the patient tumor and the corresponding PDX model, or during PDX passaging. PDXs were subjected to WES for comparison with matched patient tumor regions. Overall, our study demonstrates the feasibility of systematic multi-region PDX derivation and suggests that multiple spatial sampling of tumors could improve PDX take rates and generate PDXs that represent the intratumor diversity of heterogenous NSCLCs. Table 1. Engraftment rates of multiple spatial regions from non-small cell lung cancer tumors. Attempted NSCLC PDX B-Lymphoproliferations No Xenograft: Regions No Xenograft: Any Region LUAD 48 (21) 12 (7) 9 (7) 27 (10) LUSC 61 (17) 20 (11) 6 (5) 35 (4) Other 25 (6) 12 (4) 1 (1) 12 (1) TOTAL 134 (44) 44 (22) 16 (13) 74 (15) Citation Format: Robert E. Hynds, David R. Pearce, Ayse U. Akarca, Sophia Ward, Ariana Huebner, Oriol Pich, Gareth A. Wilson, Kate H. Gowers, Rebecca Towns, Assma Ben Aissa, Selvaraju Veeriah, Sergio A. Quezada, Mariam Jamal-Hanjani, Sam M. Janes, Nicholas McGranahan, David A. Moore, Teresa Marafioti, Charles Swanton. Multi-region patient-derived xenograft models from non-small cell lung cancer patients enrolled in lung TRACERx [abstract]. In: Proceedings of the American Association for Cancer Research Annual Meeting 2022; 2022 Apr 8-13. Philadelphia (PA): AACR; Cancer Res 2022;82(12_Suppl):Abstract nr 3096.

Robustness and accuracy evaluation of moisture prediction model for black tea withering process using hyperspectral imaging

The rapid and non-destructive detection of moisture in withering leaves is an unsolved problem because the leaves are stacked together and have random orientation. To address this issue, this study aimed to establish more robust and accurate models. The performance of front side, back side and multi-region models were compared, and the front side model showed the worst transferability. Therefore, five effective wavelength (EW) selection algorithms were combined with a successive projection algorithm (SPA) to select EWs. It was found that the shuffled frog leaping algorithm (SFLA) combined with SPA was the best method for the front side model for moisture analyses. Based on the selected EWs, the extreme learning machine (ELM) became the model with the best self-verification result. Subsequently, moisture distribution maps of withering leaves were successfully generated. Considering the processing demand of withering leaves, local region models developed based on partial least squares and the SFLA-SPA method were applied to predict the moisture of withering leaves in the local and stacked region. The results showed that the RPD, Rcv and Rp values were above 1.6, 0.870 and 0.897, respectively. These results provide a useful reference for the non-destructive detection of moisture in withering leaves.

Domestic versus foreign drivers of trade (im)balances: How robust is evidence from estimated DSGE models?

Estimated DSGE models tend to ascribe a significant and often predominant part of a country's trade balance (TB) dynamics to domestic drivers (“shocks”), suggesting foreign factors to be only of secondary importance. This paper revisits the result based on more agnostic approaches to shock transmission and using “agnostic structural disturbances”. We estimate multi-region models for Germany and Spain as countries with very distinct TB patterns since 1999. Results suggest that domestic drivers remain dominant when theory-based restrictions on shock transmission are relaxed, although the transmission of foreign shocks is strengthened.

Modeling statistical dependencies in multi-region spike train data.

Neural computations underlying cognition and behavior rely on the coordination of neural activity across multiple brain areas. Understanding how brain areas interact to process information or generate behavior is thus a central question in neuroscience. Here we provide an overview of statistical approaches for characterizing statistical dependencies in multi-region spike train recordings. We focus on two classes of models in particular: regression-based models and shared latent variable models. Regression-based models describe interactions in terms of a directed transformation of information from one region to another. Shared latent variable models, on the other hand, seek to describe interactions in terms of sources that capture common fluctuations in spiking activity across regions. We discuss the advantages and limitations of each of these approaches and future directions for the field. We intend this review to be an introduction to the statistical methods in multi-region models for computational neuroscientists and experimentalists alike.

Scaling factors for multi-region stock assessments, with an application to Indian Ocean tropical tunas

Stock assessments may be spatially structured, with regions that contain separate but linked subpopulations. In such multi-region models we must determine the relative abundances among regions. Regional scaling, which has been used since 2005 in tuna assessments, estimates the relative abundance distribution from regional catch rates and areas. We describe the method and explore potential improvements to the current practice. Supported improvements included using cell ocean areas in scaling calculations; adjusting statistical weights in the standardization model based on the density of samples; including fleet effects in the standardization model; and using a region-season interaction term in the standardization model rather than a year-season term. We also recommend the use of operational data to estimate regional scaling factors.

Comparative Study of Two Different Strategies for Determination of Soluble Solids Content of Apples From Multiple Geographical Regions by Using FT-NIR Spectroscopy

Apple is one of the most popular fresh fruits with an extensive scope of regions owing to its nutrition and sweet in flavor. There is a large difference in the composition of the fruits growing in varying regions because of the variation in the growing regions, such as temperature, soil nutrients, etc. As a result, it is of significance to decrease the impact of region variability on the measurement of soluble solids content (SSC) in apples. To lessen the impact of region variability and enhance the predictive ability of on the model, our manuscript compared the performance of the two multi-region prediction models for the estimation of SSC in apples from multiple geographical regions. One multi-region prediction model was developed by merging SSC values and spectral data of all samples from multiple regions. The other multi-region prediction model was built for the determination of SSC in combination with region discriminant, model search strategy, and single-region models. Support vector machine (SVM) was applied to establish the model for discriminating the apples from multiple geographical regions. It was found that the region discriminant model achieved great results, with the classification accuracy of 99.52%. By comparing and analyzing the two multi-region prediction models, the optimal multi-region prediction model was obtained. Finally, to decrease the irrelevant spectral information and reduce the computational cost, the multi-region SSC prediction model was optimized in combination with various spectral preprocessing methods (multiple scatter correction (MSC), standard normal variate (SNV), and first derivative (FD) correction) and variable selection methods (Monte Carlo uninformative variables elimination (MC-UVE), competitive adaptive reweighted sampling (CARS), and random frog (RF)). The overall results denoted that it was more accurate to estimate SSC in apples from the different geographical regions by using the multi-region models based on the region discriminant model in combination with SNV preprocessing algorithm and MC-UVE variable selection algorithm, and the prediction accuracy preceded the single-region models.

How large do multi-region models need to be?

Given the connectedness of most states with their neighbors, any economic analysis of changes in a state's policy needs to account for the interdependence between states. We examine in how much detail one needs to model the factor and commodity flows between states, and how much, if anything, is lost in the aggregation of neighboring states into larger regions. We develop nine dynamic multi-region general equilibrium models of the United States, with different aggregations of states (a two-region model, a 7-region model, and a full 51-region model) and different assumptions regarding intermediate inputs. We examine the same policy change with these nine models and find that all nine models suggest very similar economic effects of the policy change in the first year. Our overall conclusion is that small and highly aggregate models are not necessarily any less accurate than larger models.

Capacidade máxima de adsorção de fósforo e constante de energia em Latossolo Bruno em razão de diferentes ajustes do modelo Langmuir

A capacidade máxima de adsorção de fósforo (CMAP) é um parâmetro bastante útil para caracterizar a capacidade de adsorção de fósforo (P) do solo e, por isso, o modelo de Langmuir, que possibilita essa estimativa, é bastante difundido. Porém, se o ajuste da equação for realizado por modelos não lineares ou linearizados, ou se forem escolhidos modelos de região única ou múltiplas, nem sempre os valores estimados da CMAP e da constante de energia de ligação (k) são semelhantes. O objetivo deste trabalho foi avaliar o efeito do uso de diferentes métodos de ajuste do modelo de Langmuir sobre os valores estimados de CMAP e k. Para isso, utilizouse um único solo de alta capacidade de adsorção de P, o qual foi misturado a quantidades crescentes de areia lavada, construindo-se sistemas com capacidades de sorção crescentes, mas com a fase sólida constituída da mesma mineralogia. Foi utilizado solo do horizonte B de um Latossolo Bruno com 800 g kg-1 de argila, o qual foi misturado com areia em quantidades para obterem-se solos artificiais com 0, 200, 400, 600 e 800 g kg-1 de argila. Esses solos artificiais foram incubados por 30 dias com calcário para elevar o pH(H2O) até 6,0 e, após, foram secos em estufa e peneirados. Foram realizadas as isotermas de adsorção e os dados ajustados pelo modelo de Langmuir, usando os seguintes métodos: NLin - não linear com região única; L-1R - linearização com região única; L-2RG - linearização com duas regiões, ajuste gráfico; L-3RG - linearização com três regiões, ajuste gráfico; L-2RE linerização com duas regiões, ajuste estatístico. Os resultados evidenciaram que todos os métodos utilizados estimaram valores de CMAP proporcionais ao teor de argila dos solos e poderiam ser usados para caracterizar os solos. Contudo, quando utilizados ajustes com mais de uma região de adsorção, os valores da CMAP para a última região foram sensivelmente superiores àqueles observados após a incubação do solo com doses de P em um teste adicional. Isso indica que a CMAP da última região deve ser evitada como caracterizadora da capacidade de adsorção do solo. Conforme era esperado, os valores de k foram proporcionais aos teores de argila do solo na primeira (ou única) região dos modelos linearizados; contudo, não seguiram essa tendência no modelo não linear, recomendando-se cautela na interpretação da constante k ajustada por modelos não lineares.

New Economic Geography and Reunified Germany at Twenty: A Fruitful Match?

We qualitatively match new economic geography (NEG) to stylized facts on German economic integration after 1989. We find that NEG may explain German integration reasonably well. Germany may currently be close to the peak of the bell curve, which describes the long-run relationship between integration and agglomeration in Germany. As a consequence, further economic integration between the two parts of Germany may eventually foster redispersion of economic activity toward East Germany. We also identify limitations of NEG for explaining German integration, most notably the analytical complexity of multi-region models and its neglect of knowledge spillovers and labour pooling. RÉSUMÉ Nous établissons une correspondance qualitative entre la nouvelle géographie économique (NEG) et des faits stylisés sur l'intégration économique allemande après 1989. Nous découvrons que la NEG permet d'expliquer raisonnablement bien l'intégration allemande. En fait, l'Allemagne est sans doute proche du sommet de la courbe en cloche, décrivant les rapports de longue durée entre l'intégration et l'agglomération en Allemagne. Donc, la poursuite de l'intégration économique entre les deux parties de l'Allemagne pourra favoriser une re-dispersion de l'activité économique envers l'Allemagne de l'Est. Néantmoins, nous identifions des limitations de la MEG pour expliquer l'intégration allemande, notamment la complexité analytique des modèles multirégionaux et la négligeance des retombées des connaissances (knowledge spillovers) et le groupement de la main-d'œuvre. Extracto Emparejamos cualitativamente la Nueva Geografía Económica (NEG) con hechos estilizados de la integración económica alemana después de 1989. Descubrimos que la NEG puede explicar la integración alemana razonablemente bien. Actualmente, Alemania podría estar cerca de la cumbre de la curva de campana que describe la asociación a largo plazo entre integración y aglomeración en Alemania. En consecuencia, una mayor integración económica entre las dos partes de Alemania podría fomentar últimamente la re-dispersión de actividad económica hacia la Alemania del Este. Además identificamos limitaciones de la NEG para explicar la integración alemana, consistiendo principalmente en la complejidad analítica de los modelos multirregionales y en su preterición de los desbordamientos de conocimiento (knowledge spillovers) y de la concentración de mano de obra cualificada (labour pooling). 摘要:我们将新经济地理(NEG)和1989年德国经济统一后的典型事实进行定性比较。我们发现NEG可以很好地解释德国统一。德国现在非常接近钟形曲线的峰值, 意味着德国经济会长期呈现融合和聚集并存。因此, 东西德之间进一步的经济融合将最终促进经济活动流向东德。我们还找出了用NEG解释德国统一的局限性，尤其是多区域模型分析的复杂性以及对人才流失和劳动力储备的忽略。

Boolean operations on multi-region solids for mesh generation

An algorithm for Boolean operations on non-manifold models is proposed to allow the treatment of solids with multiple regions (internal interfaces) and degenerate portions (shells and wires), in the context of mesh generation. In a solid modeler, one of the most powerful tools to create three-dimensional objects with any level of geometric complexity is the Boolean set operators. They are intuitive and popular ways to combine solids, based on the operations applied to point sets. To assure that the resulting objects have the same dimension as the original objects, without loose or dangling parts, a regularization process is usually applied after a Boolean operation. In practice, the regularization is performed classifying the topological elements and removing internal or lower-dimensional structures. However, in many engineering applications, the adopted geometric model may contain idealized internal parts, as in the case of multi-region models, or lower-dimensional parts, as in the case of solids that contain dangling slabs that are represented as zero-thickness surfaces or wireframes in the model. Therefore, the aim of this work is the development of a generic algorithm that allows the application of the Boolean set operations in a geometric modeling environment applied to finite and boundary element mesh generation. This environment adopts a non-manifold boundary representation that considers an undefined number of topological entities (group concept), and works with objects of different dimensions and with objects not necessarily plane or polyhedral (parametric curved surfaces). Numerical examples are presented to illustrate the proposed methodology.

A Trade Algorithm for Multi-Region Models Subject to Spillover Externalities

In this paper we present an algorithm that deals with trade interactions within a multi-region model. In contrast to traditional approaches this algorithm is able to handle spillover externalities. We focus on technological spillovers which are due to capital trade. The algorithm of finding a pareto-optimal solution in an intertemporal framework is embedded in a decomposed optimization process. The paper analyzes convergence and equilibrium properties of this algorithm. In the final part of the paper, we apply the algorithm to investigate possible impacts of technological spillovers. While benefits of technological spillovers are significant for the capital-importing region, benefits for the capital-exporting region depend on the type of regional disparities and the resulting specialization and terms-of-trade effects.

The advantage of avoiding the Armington assumption in multi-region models

Most computable general equilibrium models that describe interactions between multiple regions employ the so-called Armington assumption that goods that are produced in different regions are imperfect substitutes for each other. The paper describes an alternative algorithm based on the assumption of perfect substitutability that accommodates the observation that each region produces all traded goods. For models with many regions and less-that-perfectly mobile factors, the alternative algorithm greatly reduces convergence time. The algorithm is illustrated by comparing two static multi-region models of the 50 U.S. states and the District of Columbia, with and without the Armington assumption. The model without the Armington assumption is then extended to a dynamic 51-region model with agents that have perfect foresight over 6 periods.

A modeling methodology for finite element mesh generation of multi-region models with parametric surfaces

This paper presents a description of the reorganization of a geometric modeler, MG, designed to support new capabilities of a topological module (CGC) that allows the detection of closed-off solid regions described by surface patches in non-manifold geometric models defined by NURBS. These patches are interactively created by the user by means of the modeler's graphics interface, and may result from parametric–surface intersection in which existing surface meshes are used as a support for a discrete definition of intersection curves. The geometry of realistic engineering objects is intrinsically complex, usually composed by several materials and regions. Therefore, automatic and adaptive meshing algorithms have become quite useful to increase the reliability of the procedures of a FEM numerical analysis. The present approach is concerned with two aspects of 3D FEM simulation: geometric modeling, with automatic multi-region detection, and support to automatic finite-element mesh generation.

The Common Agricultural Policy in Multisectoral Models

The paper reviews the multisectoral models used in the last 10 years to analyse the effects of the Common Agricultural Policy. It begins with a presentation of the theoretical structure of computable general equilibrium models, including both single-region and multi-region models. In this respect, the paper deals with the questions of disaggregation of commodities, households, regions and factors of production, as well as the issues of parameter specification and model closure. The paper then turns to the problems of modelling policies which are included in the present CAP. Firstly, there is the issue of the linkages between institutional and market prices. Secondly, there is the issue of decoupled or semi-decoupled direct payments. Thirdly, there are those related to monitoring the effects of the implementation of Uruguay Round Agreement commitments. Fourthly, there are the questions on the modelling of different policy instruments either related to the use of output, or input. The paper closes with an appraisal of the state of the art and recommends directions for future research, data needs and modelling efforts on the CAP.

Modeling multiphase non-isothermal fluid flow and reactive geochemical transport in variably saturated fractured rocks: 1. Methodology

Reactive fluid flow and geochemical transport in unsaturated fractured rocks have received increasing attention for studies of contaminant transport, ground- water quality, waste disposal, acid mine drainage remediation, mineral deposits, sedimentary diagenesis, and fluid-rock interactions in hydrothermal systems. This paper presents methods for modeling geochemical systems that emphasize: (1) involvement of the gas phase in addition to liquid and solid phases in fluid flow, mass transport, and chemical reactions; (2) treatment of physically and chemically heterogeneous and fractured rocks, (3) the effect of heat on fluid flow and reaction properties and processes, and (4) the kinetics of fluid-rock interaction. The physical and chemical process model is embodied in a system of partial differential equations for flow and transport, coupled to algebraic equations and ordinary differential equations for chemical interactions. For numerical solution, the continuum equations are discretized in space and time. Space discretization is based on a flexible integral finite difference approach that can use irregular gridding to model geologic structure; time is discretized fully implicitly as a first-order finite difference. Heterogeneous and fractured media are treated with a general multiple interacting continua method that includes double-porosity, dual-permeability, and multi-region models as special cases. A sequential iteration approach is used to treat the coupling between fluid flow and mass transport on the one hand, chemical reactions on the other. Applications of the methods developed here to variably saturated geochemical systems are presented in a companion paper (part 2, this issue).

A method for solving multi-region models

The paper develops a two-step method for solving multi-region general equilibrium models. The method has the advantage that the number of regions can be arbitrarily large.

Transport in chemically and mechanically heterogeneous porous media IV: large-scale mass equilibrium for solute transport with adsorption

In this article we consider the transport of an adsorbing solute in a two-region model of a chemically and mechanically heterogeneous porous medium when the condition of large-scale mechanical equilibrium is valid. Under these circumstances, a one-equation model can be used to predict the large-scale averaged velocity, but a two-equation model may be required to predict the regional velocities that are needed to accurately describe the solute transport process. If the condition of large-scale mass equilibrium is valid, the solute transport process can be represented in terms of a one-equation model and the analysis is simplified greatly. The constraints associated with the condition of large-scale mass equilibrium are developed, and when these constraints are satisfied the mass transport process can be described in terms of the large-scale average velocity, an average adsorption isotherm, and a single large-scale dispersion tensor. When the condition of large-scale mass equilibrium is not valid, two equations are required to describe the mass transfer process, and these two equations contain two adsorption isotherms, two dispersion tensors, and an exchange coefficient. The extension of the analysis to multi-region models is straight forward but tedious.

Book Review: Integrated Multiregion Models for Policy Analysis: An Australian Perspective

Book Review: Integrated Multiregion Models for Policy Analysis: An Australian Perspective

BOOK REVIEWS

A Marginal Economy? East Anglian Breckland in the Later Middle Ages, by Mark Bailey. Peripheralisation and Industrial Change, edited by G. J. R. Linge. Recursive Methods in Economic Dynamics, by Nancy L. Stokey and Robert E. Lucas, Jr., with Edward C. Prescott. Facility Location Analysis: Theory and Applications, edited by François V. Louveaux, Martine Labbé, and Jacques‐François Thisse. Expert Systems: Applications to Urban Planning, edited by Tschangho John Kim, Lyna L. Wiggins, and Jeff R. Wright. Modelling Urban Housing Market Dynamics, by L. J. G. Van Wissen and Annamarie Rima. Choice and Allocation Models for the Housing Market, by Jan Rouwendal. The Economic Value of Visibility, edited by George Tolley and Robert Fabian. Benefit‐Cost Analysis: A Political Economy Approach, by A. Allan Schmid. Regional Economic Modeling in Theory and in Practice, by Curtis C. Harris, Jr. and Mehrzad Nadji. Integrated Multiregion Models for Policy Analysis: An Australian Perspective, by Christine Smith. Information Technology: Social and Spatial Perspectives, edited by Isao Orishimo, Geoffrey J. D. Hewings, and Peter Nijkamp.