Bayesian Smoothing Research Articles

Assessment of county-level proxy variables for household firearm ownership

Our objective in this study was to evaluate how well proxy variables for firearm ownership used in county-level studies measure firearm ownership. We applied Bayesian spatial smoothing methods to calculate county-level estimates of household firearm ownership using Behavioral Risk Factor Surveillance System (BRFSS) data (2013–2018). We compared these estimates to four proxies for county-level firearm ownership: the proportion of suicides that were firearm suicides, the average of the proportion of suicides that were firearm suicides and the proportion of homicides that were firearm homicides, gun shops per capita, and federal firearm licenses per capita. U.S. counties for which BRFSS data on household firearm ownership were collected and available for release (n = 304) were included. The median (interquartile range) prevalence of household firearm ownership was 46.6% (37.2%, 56.4%). The per capita rate of federal firearm licenses was most strongly correlated with household firearm ownership (r = 0.70; 95% CI: 0.63, 0.75) followed by the proportion of suicides that were firearm suicides (r = 0.45; 95% CI: 0.36, 0.54). These correlations were stronger among counties with populations of ≥250,000 people. The per capita rate of federal firearm licenses was the best proxy variable for firearm ownership at the county level, however, a better proxy should be identified.

Autocorrelation and spatiotemporal clusters of HIV/AIDS patients in Rongchang District, Chongqing

Objective To analyze the spatiotemporal distribution and clusters characteristics of human immunodeficiency virus (HIV)/acquired immune deficiency syndrome (AIDS) patients in Rongchang District of Chongqing from 2014 to 2019, we provide reference for relevant departments to clarify the focus of AIDS prevention and control and rationally allocate medical and health resources. Methods Newly identified HIV/AIDS in Rongchang Area from 2014 to 2019 were selected from the National HIV/AIDS Comprehensive Information System. We inputted data into GeoDa 0.95i software for carry out spatial empirical bayesian smoothing and autocorrelation analysis, and into Satscan 9.3 software for statistical analysis of spatial scanning, and finally geographic information system was used for visualization. Results A total of 1 476 new cases of HIV/AIDS were reported in Rongchang District during 2014–2019. The proportion of cases over the aged ≥50 years reported in each year ranged from 58.74% to 68.92%, and increased year by year (χ 2 trend = 4.524, P = 0.033). There was a positive spatial autocorrelation in the spatial distribution of the HIV/AIDS reporting rate for each year (Moran’s I > 0, P<0.05). The areas with the most serious HIV/AIDS epidemics, including urban areas (Changyuan Street, Changzhou Street) and the neighboring Fenggao Street, Wanling Town, Zhisheng Town and other towns (streets) in the east were Rongchang AIDS-endemic areas, were identified with both the local Moran’s I index and spatiotemporal scan statistics. Conclusion Geographical differences and spatial clustering were identified in HIV/AIDS cases in Rongchang District. For the further prevention and control of the hot-spot regions expanding, the government must strengthen monitoring of the migrating population in hotspots, and strengthen AIDS interventions for middle-aged and elderly people and low-fee sex workers, while further standardize the follow-up management of AIDS in medical institutions at all levels. 摘要：目的 了解2014—2019年重庆市荣昌区艾滋病病毒 ( human immunodeficiency virus, HIV) 感染者和艾滋病 (acquired immune deficiency syndrome, AIDS) 病人(简称HIV/AIDS)的时空分布和聚类特征, 为相关部门明确艾滋病防 控重点及合理配置医疗卫生资源提供参考依据。 方法 在全国艾滋病综合防治信息系统选择2014—2019年现住址为 重庆市荣昌区的HIV/AIDS新发病例信息,导人GeoDa 0.95i软件进行空间经验贝叶斯平滑和空间自相关分析, 导人 Satscan 9.3软件进行时空扫描分析;采用地理信息系统(GIS)实现结果的可视化展示。 结果 2014—2019年, 荣昌区累 计报告HIV/AIDS新发病例1 476例, 各年报告的≥50岁病例数占全部病例的58.74% ~ 68.92%,, 并呈逐年上升趋势 (χ 2 趋势=4.524, P=0.033)。各年HIV/AIDS报告率空间分布均存在正向自相关性 (Moran’s I>0, P<0.05) 。局域空间自相 关和时空扫描统计量综合分析显示城区(昌元街道、昌州街道)和邻近的峰高街道、万灵镇、直升镇等乡镇 (街道)为荣 昌区AIDS流行地区。 结论 荣昌区HIV/AIDS新发病例在地理空间分布上存在聚集性, 各级政府应重点加强城区和邻 近城乡结合部热点区域流动人口 HIV的监测以及中老年人群和低档暗娼人群艾滋病的干预力度, 同时进一步规范各 级医疗机构艾滋病随访管理工作, 防止热点区域的扩大和HIV/AIDS的蔓延。

SARS-CoV-2 transmission among children and staff in daycare centres during a nationwide lockdown in France: a cross-sectional, multicentre, seroprevalence study

BackgroundThe extent to which very young children contribute to the transmission of SARS-CoV-2 is unclear. We aimed to estimate the seroprevalence of antibodies against SARS-CoV-2 in daycare centres that remained open for key workers' children during a nationwide lockdown in France.MethodsChildren and staff who attended one of 22 daycare centres during a nationwide lockdown in France (between March 15 and May 9, 2020) were included in this cross-sectional, multicentre, seroprevalence study. Hospital staff not occupationally exposed to patients with COVID-19, or to children, were enrolled in a comparator group. The primary outcome was SARS-CoV-2 seroprevalence in children, daycare centre staff, and the comparator group. The presence of antibodies against SARS-CoV-2 in capillary whole blood was measured with a rapid chromatographic immunoassay. We computed raw prevalence as the percentage of individuals with a positive IgG or IgM test, and used Bayesian smoothing to account for imperfect sensitivity and specificity of the assay. This study is registered with ClinicalTrials.gov, NCT04413968.FindingsBetween June 4 and July 3, 2020, we enrolled 327 children (mean age 1·9 [SD 0·9] years; range 5 months to 4·4 years), 197 daycare centre staff (mean age 40 [12] years), and 164 adults in the comparator group (42 [12] years). Positive serological tests were observed for 14 children (raw seroprevalence 4·3%; 95% CI 2·6–7·1) and 14 daycare centre staff (7·7%; 4·2–11·6). After accounting for imperfect sensitivity and specificity of the assay, we estimated that 3·7% (95% credible interval [95% CrI] 1·3–6·8) of the children and 6·8% (3·2–11·5) of daycare centre staff had SARS-CoV-2 infection. The comparator group fared similarly to the daycare centre staff; nine participants had a positive serological test (raw seroprevalence 5·5%; 95% CI 2·9–10·1), leading to a seroprevalence of 5·0% (95% CrI 1·6–9·8) after accounting for assay characteristics. An exploratory analysis suggested that seropositive children were more likely than seronegative children to have been exposed to an adult household member with laboratory-confirmed COVID-19 (six [43%] of 14 vs 19 [6%] of 307; relative risk 7·1 [95% CI 2·2–22·4]).InterpretationAccording to serological test results, the proportion of young children in our sample with SARS-CoV-2 infection was low. Intrafamily transmission seemed more plausible than transmission within daycare centres. Further epidemiological studies are needed to confirm this exploratory hypothesis.FundingAssistance Publique—Hôpitaux de Paris; Mairie de Paris, Conseil Départemental de Seine Saint Denis.TranslationsFor the French translation of the abstract see Supplementary Materials section.

Co-Occurrence of Respiratory Distress Syndrome-Associated Neonatal Mortality and Preterm Live Births: A Spatial Cluster Analysis

Background: Respiratory distress syndrome (RDS) deaths and prematurity are impacted by geographically associated antenatal and perinatal factors. This study evaluated the spatial pattern of RDS-associated neonatal mortality and the preterm live birth rates and their co-occurrences. Methods: Population-based historical cohort study of all live births from mothers living in Sao Paulo State, Brazil, during 2004-2013. Exclusion criteria: <22 weeks gestational age, <400g birthweight or congenital anomalies. RDS-associated neonatal mortality was defined as deaths <28 days with ICD-10 codes P22.0 or P28 . 0. RDS-associated neonatal mortality and preterm live birth rates were aggregated per municipality and submitted to spatial first-order (crude rates and spatial moving average) and second-order (Global Moran Indicator - I - and Local Indicator of Spatial Autocorrelation - LISA) analysis, before and after applying a smoothing method for outliers with local Bayes equation. Spearman correlation test was applied to municipalities with significant LISA for both rates. Findings: 524094 preterm live births and 9464 RDS-associated neonatal deaths were included. Global Moran Indicator suggested non-random spatial distribution of preterm live births rates before (I=0·45; p=0·001) and after Bayesian smoothing (I=0·79; p=0·001). The same was observed for RDS-associated neonatal mortality rates before (I=0·26; p=0·007) and after smoothing (I=0·71; p=0·001). LISA maps showed similar clusters for both rates, however with a negative correlation (r=-0·22; p=0·009). Interpretations: Similar spatial clusters for RDS-associated neonatal mortality and live birth rates with negative correlation suggests an unequal neonatal specialized health structure availability throughout the State, indicating areas for improvement in perinatal care. Funding Statement: We thank FAPESP for the funding. Declaration of Interests: Dr. Scavacini Marinonio reports grants from Fundacao de Amparo a Pesquisa do Estado de Sao Paulo (FAPESP), during the conduct of the study. Dr. Testoni Costa-Nobre reports grants from Fundacao de Amparo a Pesquisa do Estado de Sao Paulo (FAPESP), during the conduct of the study. Dr. Harumi Miyoshi reports grants from Fundacao de Amparo a Pesquisa do Estado de Sao Paulo (FAPESP), during the conduct of the study. Dr. Xavier Balda reports grants from Fundacao de Amparo a Pesquisa do Estado de Sao Paulo (FAPESP), during the conduct of the study. Dr. Nema Areco reports grants from Fundacao de Amparo a Pesquisa do Estado de Sao Paulo (FAPESP), during the conduct of the study. Ethics Approval Statement: The study was approved by the Ethics Committee of Universidade Federal de Sao Paulo, under the number 4.055.489, with informed consent waive, since it was a populational retrospective unidentified database.

Nowcasting for Real-Time COVID-19 Tracking in New York City: An Evaluation Using Reportable Disease Data From Early in the Pandemic.

BackgroundNowcasting approaches enhance the utility of reportable disease data for trend monitoring by correcting for delays, but implementation details affect accuracy.ObjectiveTo support real-time COVID-19 situational awareness, the New York City Department of Health and Mental Hygiene used nowcasting to account for testing and reporting delays. We conducted an evaluation to determine which implementation details would yield the most accurate estimated case counts.MethodsA time-correlated Bayesian approach called Nowcasting by Bayesian Smoothing (NobBS) was applied in real time to line lists of reportable disease surveillance data, accounting for the delay from diagnosis to reporting and the shape of the epidemic curve. We retrospectively evaluated nowcasting performance for confirmed case counts among residents diagnosed during the period from March to May 2020, a period when the median reporting delay was 2 days.ResultsNowcasts with a 2-week moving window and a negative binomial distribution had lower mean absolute error, lower relative root mean square error, and higher 95% prediction interval coverage than nowcasts conducted with a 3-week moving window or with a Poisson distribution. Nowcasts conducted toward the end of the week outperformed nowcasts performed earlier in the week, given fewer patients diagnosed on weekends and lack of day-of-week adjustments. When estimating case counts for weekdays only, metrics were similar across days when the nowcasts were conducted, with Mondays having the lowest mean absolute error of 183 cases in the context of an average daily weekday case count of 2914.ConclusionsNowcasting using NobBS can effectively support COVID-19 trend monitoring. Accounting for overdispersion, shortening the moving window, and suppressing diagnoses on weekends—when fewer patients submitted specimens for testing—improved the accuracy of estimated case counts. Nowcasting ensured that recent decreases in observed case counts were not overinterpreted as true declines and supported officials in anticipating the magnitude and timing of hospitalizations and deaths and allocating resources geographically.

Stable Non-Linear Generalized Bayesian Joint Models for Survival-Longitudinal Data

Joint models have received increasing attention during recent years with extensions into various directions; numerous hazard functions, different association structures, linear and non-linear longitudinal trajectories amongst others. Here, we present a partially linear joint model with a Bayesian smoothing spline component to capture non-linear longitudinal trajectories where the longitudinal data can assume non-Gaussian distributions. Our approach is stable with regards to the knot set as opposed to most well-known spline models. We implement this method using the R-INLA package and show that most joint models with shared Gaussian random effects are part of the class of latent Gaussian models (LGMs). We present an illustrative example to show the use of the stable partially linear joint model and an application to real data using a skew-normal partially linear joint model. This paves the way for efficient implementation of joint models with various complex model components using the same methodology and computational platform.

Geospatial analysis of malaria mortality in the kintampo health and demographic surveillance area of central Ghana

ABSTRACT Malaria remains a menace to the existence of humanity in most contexts. Geospatial analysis of malaria mortality is crucial to identifying clusters of high disease burden and areas with limited access to malaria care for targeted control and remedial interventions. This study identified spatial and space-time clusters of malaria mortality in the Kintampo area of central Ghana. We used 1301 malaria deaths archived from 2005 to 2017 and Global Positioning System (GPS) point locations of the sub-districts in which these deaths occurred for our analysis. Mortality risks were smoothed and mapped using the Spatial Empirical Bayesian smoothing technique in Geoda (version 1.12.1.161) whereas spatial and spatio-temporal clustering analysis was done using SaTScan (version 9.6). Malaria mortality risks ranged between 1.2 and 2.4 deaths per 1000 population for persons of all ages and between 3.3 and 6.0 deaths per 1000 population for children under five years of age by sub-district. Two spatial clusters were detected for all-age malaria mortality with only the primary cluster (RR = 1.42, p = 0.001) being statistically significant. Also, two statistically significant space-time clusters were detected for all-age malaria mortality in the study area. The most likely cluster occurred between 2006 and 2011 in five sub-districts with a relative risk of 2.12 (p < 0.001) whilst the secondary cluster which had a relative risk of 2.47 (p < 0.001) occurred between 2005 and 2010 in four sub-districts. Similarly, only the most likely spatial cluster of under-five malaria mortality was statistically significant (RR = 1.36, p = 0.024). Furthermore, three spatio-temporal clusters of under-five malaria mortality were detected in the study area. The primary and second secondary clusters were statistically significant whilst the first secondary cluster had borderline significance. The primary cluster (RR = 4.49, p = 0.002) occurred in two sub-districts between 2006 and 2007. The first secondary cluster (RR = 2.21, P = 0.005) covered four sub-districts and was detected between 2006 and 2011 whereas the second secondary cluster (RR = 2.51, p = 0.003) covered two sub-districts between 2008 and 2013. Ultimately, our analysis identified a number of substantial spatial and apace-time clusters of malaria mortality in the study context, which could aid in the strategic planning, implementation and monitoring of targeted malaria control interventions.

Geospatial clustering of childhood IgA vasculitis and IgA vasculitis-associated nephritis

ObjectivesResearch on spatial variability of the incidence of IgA vasculitis (IgAV) in children and its potential implications for elucidation of the multifactorial aetiology and pathogenesis is limited. We intended to...

Pattern discovery of health curves using an ordered probit model with Bayesian smoothing and functional principal component analysis.

This article is motivated by the need for discovering patterns of patients' health based on their daily settings of care to aid the health policy-makers to improve the effectiveness of distributing funding for health services. The hidden process of one's health status is assumed to be a continuous smooth function, called the health curve, ranging from perfectly healthy to dead. The health curves are linked to the categorical setting of care using an ordered probit model and are inferred through Bayesian smoothing. The challenges include the nontrivial constraints on the lower bound of the health status (death) and on the model parameters to ensure model identifiability. We use the Markov chain Monte Carlo method to estimate the parameters and health curves. The functional principal component analysis is applied to the patients' estimated health curves to discover common health patterns. The proposed method is demonstrated through an application to patients hospitalized from strokes in Ontario. Whilst this paper focuses on the method's application to a health care problem, the proposed model and its implementation have the potential to be applied to many application domains in which the response variable is ordinal and there is a hidden process. Our implementation is available at https://github.com/liangliangwangsfu/healthCurveCode.

The application of spatial empirical Bayesian smoothing method in spatial analysis of bacillary dysentery: A case study in Yudu County, Jiangxi Province

Bacillary dysentery (BD) has become one of the major public health threats to the sustainable development of human beings. The primary objective of this paper is to validate the effectiveness of applying spatial empirical Bayesian smoothing (SEBS) method in spatial analysis of BD in Yudu County, Jiangxi Province. By using SEBS method in calculating BD incidence, we compared the raw incidence data and smoothed incidence data. Spatial distribution map and global autocorrelation analysis were adopted to explore the effect of SEBS method in BD analysis. The result showed that SEBS method provided a stable incidence estimate in epidemic research. The proposed method could be used as an effective tool for studying the spatial distribution of BD, especially in town level study.

Performance Analysis of Bayesian Filtering and Smoothing Algorithms for Underwater Passive Target Tracking

Passive target tracking deals with nonlinear filtering in which dynamics of the system are considered to be linear, while the target state is built on nonlinear measurements. In this paper, a comparative study is conducted for accurate state estimation of an underwater far-field moving target by exploiting the strength of well known nonlinear variant of Bayesian filter, i.e., extended Kalman filter (EKF) with discrete-time Kalman smoother, called Rauch–Tung–Striebel (RTS) smoother. Analysis is performed with two key parameters in target tracking by mean of variation in the number of sensors and different standard deviations of measured noise in the context of underwater bearings-only tracking technology. Exhaustive experiments are performed for finding the least root-mean square error between true and estimated position of target movement in the trajectory at every time instant. Relatively accurate estimation of the target state is observed from noisy measurements of sensors in case of RTS smoother than that of EKF for each scenario.

Efficient Bayesian PARCOR approaches for dynamic modeling of multivariate time series

A Bayesian lattice filtering and smoothing approach is proposed for fast and accurate modeling and inference in multivariate non‐stationary time series. This approach offers computational feasibility and interpretable time‐frequency analysis in the multivariate context. The proposed framework allows us to obtain posterior estimates of the time‐varying spectral densities of individual time series components, as well as posterior measurements of the time‐frequency relationships across multiple components, such as time‐varying coherence and partial coherence. The proposed formulation considers multivariate dynamic linear models (MDLMs) on the forward and backward time‐varying partial autocorrelation coefficients (TV‐VPARCOR). Computationally expensive schemes for posterior inference on the multivariate dynamic PARCOR model are avoided using approximations in the MDLM context. Approximate inference on the corresponding time‐varying vector autoregressive (TV‐VAR) coefficients is obtained via Whittle's algorithm. A key aspect of the proposed TV‐VPARCOR representations is that they are of lower dimension, and therefore more efficient, than TV‐VAR representations. The performance of the TV‐VPARCOR models is illustrated in simulation studies and in the analysis of multivariate non‐stationary temporal data arising in neuroscience and environmental applications. Model performance is evaluated using goodness‐of‐fit measurements in the time‐frequency domain and also by assessing the quality of short‐term forecasting.

Spatial Trends in Salmonella Infection in Pigs in Spain.

Salmonella is one of the most important foodborne pathogens worldwide. Its main reservoirs are poultry and pigs, in which infection is endemic in many countries. Spain has one of the largest pig populations in the world. Even though Salmonella infection is commonly detected in pig farms, its spatial distribution at the national level is poorly understood. Here we aimed to report the spatial distribution of Salmonella-positive pig farms in Spain and investigate the presence of potential spatial trends over a 17-year period. For this, data on samples from pigs tested for Salmonella in 2002–2013, 2015, 2017, and 2019 as part of the Spanish Veterinary Antimicrobial Resistance Surveillance program, representing 3,730 farms were analyzed. The spatial distribution and clustering of Salmonella-positive pig farms at the province level were explored using spatial empirical Bayesian smoothing and global Moran's I, local Moran's I, and the Poisson model of the spatial scan statistics. Bayesian spatial regression using a reparameterized Besag-York-Mollié Poisson model (BYM2 model) was then performed to quantify the presence of spatially structured and unstructured effects while accounting for the effect of potential risk factors for Salmonella infection at the province level. The overall proportion of Salmonella-positive farms was 37.8% (95% confidence interval: 36.2–39.4). Clusters of positive farms were detected in the East and Northeast of Spain. The Bayesian spatial regression revealed a West-to-East increase in the risk of Salmonella infection at the province level, with 65.2% (50% highest density interval: 70–100.0%) of this spatial pattern being explained by the spatially structured component. Our results demonstrate the existence of a spatial variation in the risk of Salmonella infection in pig farms at the province level in Spain. This information can help to optimize risk-based Salmonella surveillance programs in Spain, although further research to identify farm-level factors explaining this pattern are needed.

Inferring seismic hazard in Sichuan-Yunnan region of China based on the modern earthquake catalogue (1980-2019)*

Based on the modern earthquake catalogue, the incomplete centroidal voronoi tessellation (ICVT) method was used in this study to estimate the seismic hazard in Sichuan-Yunnan region of China. We calculated spatial distributions of the total seismic hazard and background seismic hazard in this area. The Bayesian delaunay tessellation smoothing method put forward by Ogata was used to calculate the spatial distributions of b-value. The results show that seismic hazards in Sichuan-Yunnan region are high, and areas with relatively high hazard values are distributed along the main faults, while seismic hazards in Sichuan basin are relatively low.

P1638EARLY RENAL FUNCTION TRAJECTORIES, CYTOMEGALOVIRUS SEROSTATUS AND LONG-TERM GRAFT OUTCOMES IN KIDNEY TRANSPLANT RECIPIENTS: A BAYESIAN ANALYSIS STUDY

Abstract Background and Aims Despite significant improvements in short-term kidney transplant survival, long-term graft survival has not improved to the same degree with transplant failure being a top four cause of end-stage renal disease. We previously showed in a prevalent kidney transplant population that most patients do not experience linear renal function trajectories1. Many, instead, have periods of stability whilst others experience rapid progression. We also showed that episodes of rapid progression are associated with graft loss. Understanding trajectories of kidney allograft function is, therefore, key to defining the mechanisms underpinning allograft dysfunction. In this study, we evaluated the allograft function trajectories and associated factors, in an unselected, incident population of kidney allograft recipients in the early period post-transplantation. We also investigate whether episodes of rapid progression or non-progression are associated with graft loss in an extended follow-up period Method Demographic and clinical data were obtained from electronic health records. We used Bayesian smoothing techniques1 to create 10,000 Monte Carlo sample curves for 310 kidney transplant recipients for estimated glomerular filtration rates from 3-27 months after transplantation. This technique produces a smooth curve for each patient that reflects the gradual, longer term changes in eGFR values, rather than the rapid, short-term changes because of clinical and biologic variation as well as other interference including measurement error. The estimated trajectory is a smooth curve, allowing its slope to be calculated month by month. The probability of having an episode of rapid progression (decline greater than 5 ml/min/1.73m2/year in any 1-month period) and non-progression (decline no greater than -1ml/min/1.73m2/year) were calculated. Overall follow-up period was 8 years. Factors associated with having an episode of rapid progression, non-progression, and associations with long-term graft loss were explored. Results A median of 54 eGFR measurements per patient were available from 3-27 months for analysis. 65 patients (21%) had a probability of rapid progression greater than 0.8. During the follow-up period, 34 patients (11%) lost their graft. In multivariable Cox Proportional Hazard analysis, a probability greater than 0.8 of rapid progression was associated with long-term death-censored graft loss (Hazard ratio, 2.17; 95% CI, 1.04-4.55). In separate multivariable logistic regression models, cytomegalovirus serostatus donor positive to recipient positive (Odds ratio [OR], 3.82; 95% CI 1.63-8.97), CMV donor positive (OR 2.06; 95% CI 1.15-3.68), and CMV recipient positive (OR 2.03; 95% CI 1.14-3.60) were associated with having a greater than 0.8 probability of an episode of rapid progression. Having a probability greater than 0.8 of non-progression was not associated graft loss. Conclusion Early episodes of rapid progression are associated with long-term death-censored graft loss and are associated with cytomegalovirus seropositivity. Possible mechanisms include adverse cytomegalovirus-related immunomodulatory effects resulting in increased infections, glomerular injury and allograft vasculopathy. Further investigation into these factors may yield potentially modifiable risk factors and improve graft survival.

Nowcasting by Bayesian Smoothing: A flexible, generalizable model for real-time epidemic tracking.

Achieving accurate, real-time estimates of disease activity is challenged by delays in case reporting. “Nowcast” approaches attempt to estimate the complete case counts for a given reporting date, using a time series of case reports that is known to be incomplete due to reporting delays. Modeling the reporting delay distribution is a common feature of nowcast approaches. However, many nowcast approaches ignore a crucial feature of infectious disease transmission—that future cases are intrinsically linked to past reported cases—and are optimized to one or two applications, which may limit generalizability. Here, we present a Bayesian approach, NobBS (Nowcasting by Bayesian Smoothing) capable of producing smooth and accurate nowcasts in multiple disease settings. We test NobBS on dengue in Puerto Rico and influenza-like illness (ILI) in the United States to examine performance and robustness across settings exhibiting a range of common reporting delay characteristics (from stable to time-varying), and compare this approach with a published nowcasting software package while investigating the features of each approach that contribute to good or poor performance. We show that introducing a temporal relationship between cases considerably improves performance when the reporting delay distribution is time-varying, and we identify trade-offs in the role of moving windows to accurately capture changes in the delay. We present software implementing this new approach (R package “NobBS”) for widespread application and provide practical guidance on implementation.

Bayesian Spatial Kernel Smoothing for Scalable Dense Semantic Mapping

This paper develops a Bayesian continuous 3D semantic occupancy map from noisy point clouds by generalizing the Bayesian kernel inference model for building occupancy maps, a binary problem, to semantic maps, a multi-class problem. The proposed method provides a unified probabilistic model for both occupancy and semantic probabilities and nicely reverts to the original occupancy mapping framework when only one occupied class exists in obtained measurements. The Bayesian spatial kernel inference relaxes the independent grid assumption and brings smoothness and continuity to the map inference, enabling to exploit local correlations present in the environment and increasing the performance. The accompanying software uses multi-threading and vectorization, and runs at about 2 Hz on a laptop CPU. Evaluations using multiple sequences of stereo camera and LiDAR datasets show that the proposed method consistently outperforms current baselines. We also present a qualitative evaluation using data collected with a bipedal robot platform on the University of Michigan - North Campus.

Software application profile: the Rapid Inquiry Facility 4.0: an open access tool for environmental public health tracking.

The Rapid Inquiry Facility 4.0 (RIF) is a new user-friendly and open-access tool, developed by the UK Small Area Health Statistics Unit (SAHSU), to facilitate environment public health tracking (EPHT) or surveillance (EPHS). The RIF is designed to help public health professionals and academics to rapidly perform exploratory investigations of health and environmental data at the small-area level (e.g. postcode or detailed census areas) in order to identify unusual signals, such as disease clusters and potential environmental hazards, whether localized (e.g. industrial site) or widespread (e.g. air and noise pollution). The RIF allows the use of advanced disease mapping methods, including Bayesian small-area smoothing and complex risk analysis functionalities, while accounting for confounders. The RIF could be particularly useful to monitor spatio-temporal trends in mortality and morbidity associated with cardiovascular diseases, cancers, diabetes and chronic lung diseases, or to conduct local or national studies on air pollution, flooding, low-magnetic fields or nuclear power plants.

Variational Bayesian Cubature RTS Smoothing for Transfer Alignment of DPOS

Multi-task remote sensing sensors have become attractive development directions of aerial remote sensing system, which rely on distributed position and orientation system (DPOS) to provide multi-node motion parameters to achieve superior performance. DPOS depends on transfer alignment from its master system to slave inertial measurement units to obtain multi-node motion information. However, for DPOS, there are many factors like carrier maneuver mode and external disturbance which will result in time-varying measurement noise and further degrade transfer alignment performance of DPOS obviously. In this work, a transfer alignment method based on variational Bayesian cubature RTS smoothing is developed to improve the accuracy of DPOS, which is implemented by combining the cubature RTS smoothing algorithm and variational Bayesian estimation method to deal with the time-varying measurement noise. A semi-physical simulation based on real flight experiment has been conducted, the results show that the motion parameter accuracy has achieved noticeable enhancement than the existing cubature RTS smoothing algorithm.

Temporal Parallelization of Bayesian Smoothers

This article presents algorithms for temporal parallelization of Bayesian smoothers. We define the elements and the operators to pose these problems as the solutions to all-prefix-sums operations for which efficient parallel scan-algorithms are available. We present the temporal parallelization of the general Bayesian filtering and smoothing equations, and specialize them to linear/Gaussian models. The advantage of the proposed algorithms is that they reduce the linear complexity of standard smoothing algorithms with respect to time to logarithmic.