Parameter Estimation Research Articles

Comparing sensitivities of groundwater head to variations in hydraulic parameters under boundary conditions of river stage rise and tidal variation

The sensitivity of hydraulic head responses to spatially distributed hydraulic parameters is essential for uncertainty analysis, inverse modeling, and parameter estimation and interpretation. This study formulates the Fréchet sensitivity kernel of hydraulic head responses to a suddenly rising boundary and a sinusoidal head fluctuation boundary to variation of spatially distributed hydraulic parameters in a semi-infinite, one-dimensional (1-D), confined aquifer, and it then derives analytical solutions. Different from previous studies that derived expressions for Fréchet kernels in the time domain for a 2-D pumping test, this study is the first to derive the closed-form Fréchet kernels in time and frequency domains for a semi-infinite, 1-D, confined aquifer. This study uses the Fréchet kernels to investigate the nature of singularities in the spatial sensitivity functions around the observation location and boundary. The information content revealed by observation of head change or head fluctuation amplitude at a given specified location and time (or frequency) under the above two boundary conditions is different. When comparing Fréchet sensitivity kernels across various times or periods, multi-frequency information, much like multi-time information, can be instrumental for hydrogeological parameter inversion. The explicit-form Fréchet sensitivity kernels also identify the optimal time or period for obtaining measurements.

Fast Bayesian Functional Regression for Non-Gaussian Spatial Data

Functional generalized linear models (FGLM) have been widely used to study the relations between non-Gaussian response and functional covariates. However, most existing works for FGLM assume independence among observations and therefore they are of limited applicability for correlated data. A particularly important example is spatial functional data, where we observe functions over spatial domains, such as the age population curve or temperature curve at each areal unit. In this paper, we extend FGLM by incorporating spatial random effects. Especially, we study the relationship between the non-Gaussian response variable and functional covariates that are spatially observed. However, such model has computational and inferential challenges. The high-dimensional spatial random effects cause the slow mixing of Markov chain Monte Carlo (MCMC) algorithms. Furthermore, spatial confounding can lead to bias in parameter estimates and inflate their variances. To address these issues, we propose an efficient Bayesian method using a sparse reparameterization of high-dimensional random effects. We also study the average coverage probabilities of the credible intervals of functional parameters. We apply our methods to simulated and real data examples, including malaria incidence data and US COVID-19 data. The proposed method is fast while providing accurate functional estimates.

Genetic parameters for reproductive traits and somatic cell score in U.S. dairy goats.

There are no studies regarding the estimation of genetic parameters and genetic trends for reproductive traits and somatic cells in goats. Their knowledge allows optimization of selection schemes. The objective of this study was to estimate genetic parameters and genetic and phenotypic trends for age at first kidding (AFK), kidding interval (KIN) and somatic cell score (SCS). Analyses were conducted within and across seven US goat breeds, namely, Nubian (NU), Alpine (AL), LaMancha (LM), Toggenburg (TO), Saanen (SA), Nigerian Dwarf (ND) and Oberhasli (OB), and a set of all of these breeds (AB). The restricted maximum likelihood methodology and trivariate animal models were used. Genetic and phenotypic trends were estimated using regression models. The average and standard deviation of AFK, KIN and SCS for AB were 573.6 ± 178.5days, 418.8 ± 125.5days and 4.67 ± 2.23 Log2, respectively. The heritabilities (h2) and standard errors of AFK, KIN and SCS for AB were 0.28 ± 0.02, 0.04 ± 0.02 and 0.22 ± 0.01, respectively. The h2 ranged from 0.15 (SA) to 0.37 (NU) for AFK, from 0.04 (AB) to 0.10 (AL) for KIN, and from 0.11 (TO) to 0.26 (LM and ND) for SCS. Genetic correlations between AFK and KIN and between AFK and SCS for AB were positive and weak (0.07 and 0.12, respectively) but significant (P < 0.01). Genetic correlations between SCS and KIN were significant (P < 0.01) for all the breeds and ranged from -0.15 (NU) to 0.44 (AL). Genetic correlations between AFK and SCS in the NU and AL breeds were similar (approximately 0.21). A positive genetic trend was found for KIN in the SA breed, which caused an increase in the number of days between consecutive kiddings. The genetic trend of SCS for the NU, AL and ND breeds was negative and decreased annually, which is beneficial for producers. These first results show the intensity and direction of some favorable/unfavorable relationships between AFK or KIN and SCS Log2 in some U.S. goat genetic groups.

A mixture distribution for modelling bivariate ordinal data

AbstractOrdinal responses often arise from surveys which require respondents to rate items on a Likert scale. Since most surveys contain more than one question, the data collected are multivariate in nature, and the associations between different survey items are usually of considerable interest. In this paper, we focus on a mixture distribution, called the combination of uniform and binomial (CUB), under which each response is assumed to originate from either the respondent’s uncertainty or the actual feeling towards the survey item. We extend the CUB model to the bivariate case for modelling two correlated ordinal data without using copula-based approaches. The proposed model allows the associations between the unobserved uncertainty and feeling components of the variables to be estimated, a distinctive feature compared to previous attempts. This article describes the underlying logic and deals with both theoretical and practical aspects of the proposed model. In particular, we will show that the model is identifiable under a wide range of conditions. Practical inferential aspects such as parameter estimation, standard error calculations and hypothesis tests will be discussed through simulations and a real case study.

Optimal parameter estimation for uncertain structural systems under unknown random excitations

An optimal parameter estimation method for uncertain structural systems under unknown random excitations is proposed, which combines the Bayesian inference and stochastic dynamics using geometrically averaged likelihood and optimal response estimation. The general description of optimal parameter estimation problems for uncertain dynamic systems under only stochastic responses measured is presented. The posterior probability density conditional on measured responses is expressed by the likelihood function conditional on system parameters based on Bayes’ theorem. For finite time processes, the parameter estimation problem as the probability integral of conditional means is converted into the optimization problem expressed as maximizing the posterior probability density. A geometrically averaged likelihood function is defined and used for calculating the logarithmic posterior probability density. This estimation can avoid the numerical singularity of the likelihood function and reduce the effects of incomplete posterior probability density and inaccurate prior statistics of unknown random excitations, and then it will be more reasonable and effective. Furthermore, the differential equations for system response means and covariances are derived and solved based on stochastic dynamics theory. The means and covariances conditional on responses at the present instant are expressed by the previous statistics based on optimal response estimation. By combining two results, the analytical expressions of the averaged likelihood function and logarithmic posterior probability density are obtained which will be more reliable and accurate. The proposed optimal estimation method is verified by numerical results for a five-storey frame structure under base random excitation. The estimated results are not affected by the prior statistics errors of random excitation as a factor. For noisy observation, the Kalman filtering is incorporated in the estimation method and hence the estimation is more accurate and robust even for low signal noise ratios. The optimal estimation method has the potential for application to general uncertain systems.

A New Three-Parameter Inverse Rayleigh Distribution: Simulation and Application to Real Data

In this paper, we introduce a new three-parameter inverse Rayleigh distribution that extends the inverse Rayleigh distribution, constructed based on the generalized transmuted family of distributions proposed by Alizadeh, Merovci, and Hamedani. We explore statistical properties such as the quantile function, moments, harmonic mean, mean deviation, stress–strength reliability, and entropy. Parameter estimation is performed using various methods, including maximum likelihood, least squares, the method of the maximum product of spacings, and the method of Cramér–von Mises. The usefulness of the new three-parameter inverse Rayleigh distribution is illustrated by modeling a real dataset, demonstrating its superior fit compared to several other distributions.

Random forest-based multipath parameter estimation

Multipath is recognized as one of the major error sources for GNSS urban navigation. This study proposes a random forest (RF)-based multipath parameter estimator that uses random forest regression for parameter estimation, thereby mitigating multipath effect by removing the estimated reflected signal components. The proposed estimator is evaluated and compared with the multipath estimation delay-lock loop (MEDLL) for one-multipath and three-multipath cases, respectively. Simulation results demonstrate that the RF-based estimator is less affected by the front-end bandwidth of received signals, compared with MEDLL. The proposed RF-based estimator shows better performance than MEDLL for signals with front-end bandwidths of lower than 6 MHz. In 20 sets of tests on signals with a front-end bandwidth of 10 MHz in the three-multipath case, the RF-based estimator obtains smaller standard deviations than MEDLL. In experiments using real data with a front-end bandwidth of 2 MHz, the RF-based estimator reduces the 2D and 3D positioning errors by 8.5% and 8.7% over 180 epochs, respectively, against the conventional delayed-locked loop (DLL).

Reliability inference for a family of inverted exponentiated distributions under block progressive Type II censoring

In this article, estimation of survival characteristics is discussed when life tests are performed against different test facilities using block progressive Type-II censoring. Under the assumption that the lifetimes of test units follow a family of inverted exponentiated distributions, estimates are derived under classical and hierarchical Bayesian frameworks when the differences in different test facilities are taken into consideration. Maximum likelihood estimators are obtained and their existence and uniqueness properties are also established. In addition, Bayes estimators are provided under a hierarchical framework, and a hybrid Metropolis-Hastings sampling procedure is proposed for complex posterior computation. The performance of all the proposed approaches is compared via extensive simulation studies and two real-life data are analyzed for illustration. From numerical results we observe that the hierarchical approach provides relatively better estimates for model parameters.

Using machine learning techniques in inverse problems of acoustical oceanography

AbstractThe goal of the work presented here is to study a novel approach for inverting acoustic signals recorded in the marine environment for the estimation of environmental parameters of the water column and/or the seabed. The proposed approach is based on signal feature extraction using a discrete wavelet packet transform, applied to the measured signal, and hidden Markov models that exploit the sequential patterns of the signals. The signal feature is thereafter used in the framework of a mixture density network, which, after training with sets of simulated signals calculated within a predefined search space, provides conditional posterior distributions of the recoverable parameters. The technique is tested with two test cases corresponding to different types of inverse problems. The first case corresponds to a simple problem of geoacoustic inversion, while the second is referred to a, rather unusual, still interesting problem of recovering the shape of a seamount using long‐range acoustic data. Both test cases are based on simulated experiments. The inversion results obtained using the proposed scheme are compared with inversion results using statistical features of the acoustic signal, which is another inversion approach well documented in the literature and is also based on the wavelet packet transform of the measured signal.

FlexDTI: flexible diffusion gradient encoding scheme-based highly efficient diffusion tensor imaging using deep learning

Objective. Most deep neural network-based diffusion tensor imaging methods require the diffusion gradients’ number and directions in the data to be reconstructed to match those in the training data. This work aims to develop and evaluate a novel dynamic-convolution-based method called FlexDTI for highly efficient diffusion tensor reconstruction with flexible diffusion encoding gradient scheme. Approach. FlexDTI was developed to achieve high-quality DTI parametric mapping with flexible number and directions of diffusion encoding gradients. The method used dynamic convolution kernels to embed diffusion gradient direction information into feature maps of the corresponding diffusion signal. Furthermore, it realized the generalization of a flexible number of diffusion gradient directions by setting the maximum number of input channels of the network. The network was trained and tested using datasets from the Human Connectome Project and local hospitals. Results from FlexDTI and other advanced tensor parameter estimation methods were compared. Main results. Compared to other methods, FlexDTI successfully achieves high-quality diffusion tensor-derived parameters even if the number and directions of diffusion encoding gradients change. It reduces normalized root mean squared error by about 50% on fractional anisotropy and 15% on mean diffusivity, compared with the state-of-the-art deep learning method with flexible diffusion encoding gradient scheme. Significance. FlexDTI can well learn diffusion gradient direction information to achieve generalized DTI reconstruction with flexible diffusion gradient scheme. Both flexibility and reconstruction quality can be taken into account in this network.

Cost-effectiveness analysis of surgical masks, N95 masks compared to wearing no mask for the prevention of COVID-19 among health care workers: Evidence from the public health care setting in India.

Nonpharmacological interventions, such as personal protective equipment for example, surgical masks and respirators, and maintenance of hand hygiene along with COVID-19 vaccines have been recommended to reduce viral transmission in the community and health care settings. There is evidence from the literature that surgical and N95 masks may reduce the initial degree of exposure to the virus. A limited research that has studied the cost-effective analysis of surgical masks and N95 masks among health care workers in the prevention of COVID-19 in India. The objective of this study was to estimate the cost-effectiveness of N95 and surgical mask compared to wearing no mask in public hospital settings for preventing COVID-19 infection among Health care workers (HCWs) from the health care provider's perspective. A deterministic baseline model, without any mask use, based on Eikenberry et al was used to form the foundation for parameter estimation and to estimate transmission rates among HCWs. Information on mask efficacy, including the overall filtering efficiency of a mask and clinical efficiency, in terms of either inward efficiency(ei) or outward efficiency(e0), was obtained from published literature. Hospitalized HCWs were assumed to be in one of the disease states i.e., mild, moderate, severe, or critical. A total of 10,000 HCWs was considered as representative of the size of a tertiary care institution HCW population. The utility values for the mild, moderate and severe model health states were sourced from the primary data collection on quality-of-life of HCWs COVID-19 survivors. The utility scores for mild, moderate, and severe COVID-19 conditions were 0.88, 0.738 and 0.58, respectively. The cost of treatment for mild sickness (6,500 INR per day), moderate sickness (10,000 INR per day), severe (require ICU facility without ventilation, 15,000 INR per day), and critical (require ICU facility with ventilation per day, 18,000 INR) per day as per government and private COVID-19 treatment costs and capping were considered. One way sensitivity analyses were performed to identify the model inputs which had the largest impact on model results. The use of N95 masks compared to using no mask is cost-saving of $1,454,632 (INR 0.106 billion) per 10,000 HCWs in a year. The use of N95 masks compared to using surgical masks is cost-saving of $63,919 (INR 0.005 billion) per 10,000 HCWs in a year. the use of surgical masks compared to using no mask is cost-saving of $1,390,713 (INR 0.102 billion) per 10,000 HCWs in a year. The uncertainty analysis showed that considering fixed transmission rate (1.7), adoption of mask efficiency as 20%, 50% and 80% reduces the cumulative relative mortality to 41%, 79% and 94% respectively. On considering ei = e0 (99%) for N95 and surgical mask with ei = e0 (90%) the cumulative relative mortality was reduced by 97% and the use of N95 masks compared to using surgical masks is cost-saving of $24,361 (INR 0.002 billion) per 10,000 HCWs in a year. Both considered interventions were dominant compared to no mask based on the model estimates. N95 masks were also dominant compared to surgical masks.

High-Speed Surface Property Recognition with a 140 GHz Frequency

In the field of integrated sensing and communication, there is a growing need for advanced environmental perception. The terahertz (THz) frequency band, significant for ultra-high-speed data connections, shows promise in environmental sensing, particularly in detecting surface textures crucial for autonomous systems’ decision-making. However, traditional numerical methods for parameter estimation in these environments struggle with accuracy, speed, and stability, especially in high-speed scenarios like vehicle-to-everything communications. This study introduces a deep learning approach for identifying surface roughness using a 140-GHz setup tailored for such conditions. A high-speed data acquisition system was developed to mimic real-world scenarios, and a diverse set of rough surface samples was prepared for realistic high-speed datasets to train the models. The model was trained and validated in three challenging scenarios: random occlusions, sparse data, and narrow-angle observations. The results demonstrate the method’s effectiveness in high-speed conditions, suggesting terahertz frequencies’ potential in future sensing and communication applications.

Intraoperative Ventilation/Perfusion Mismatch and Postoperative Pulmonary Complications after Major non-cardiac surgery: a prospective cohort study.

Postoperative pulmonary complications (PPCs) can increase hospital length of stay, postoperative morbidity and mortality. Despite many factors can increase the risk of PPCs, it is not known whether intraoperative ventilation/perfusion (V/Q) mismatch can be associated with an increased risk of PPCs after major non-cardiac surgery. We enrolled patients undergoing general anesthesia for non-cardiac surgery and evaluated intraoperative V/Q distribution using the Automatic Lung Parameter Estimator technique. The assessment was done after anesthesia induction (T1), after 1 hour from surgery start (T2) and at the end of surgery (T3). We collected demographic and procedural information and measured intraoperative ventilatory and hemodynamic parameters at each time-point. Patients were followed up for 7 days after surgery and assessed daily for PPCs occurrence. We enrolled 101 patients with a median age of 71 [62-77] years, a BMI of 25 [22.4-27.9] kg/m 2 and a preoperative ARISCAT score of 41 [34-47]. Of them, 29 (29%) developed PPCs, mainly acute respiratory failure (23%) and pleural effusion (11%). Patients with and without PPCs did not differ in levels of shunt at T1 (PPCs:22.4[10.4-35.9] % vs No PPCs:19.3[9.4-24.1] %, p=0.18) or during the protocol, while significantly different levels of high V/Q were found during surgery (PPCs:13[11-15] mmHg vs No PPCs:10[8-13.5] mmHg, p=0.007) and before extubation (PPCs:13[11-14]mmHg vs No PPCs:10[8-12] mmHg, p=0.006). After adjusting for age, ARISCAT, BMI, smoking, fluid balance, anesthesia type, laparoscopic procedure and surgery duration, high V/Q before extubation was independently associated with the development of PPCs (OR 1.147, CI 95% [1.021-1.289], p=0.02). The sensitivity analysis showed an E-value of 1.35 (CI=1.11). In patients with intermediate/high risk of PPCs undergoing major non-cardiac surgery, intraoperative V/Q mismatch is associated with the development of PPCs. Increased high V/Q before extubation is independently associated with the occurrence of PPCs in the first 7 days after surgery.

Predictive Partly Conditional Models for Longitudinal Ordinal Outcomes with Application to Alzheimer’s Disease Progression

AbstractAssessing time-dependent risk factors in relation to the risk of disease progression is challenging, yet important, especially for chronic diseases with slow progression. In this paper, the partly conditional model is extended for characterizing disease progression at time t with longitudinal ordinal outcomes in the presence of time-dependent covariates at time s$$(s &lt; t)$$ ( s &lt; t ) and time-dependent effects. Advantages of the method include direct modeling of disease progression, use of longitudinal risk factors, and flexibility in target period of progression $$u = t - s$$ u = t - s . A generalized estimating equation approach is adopted for parameter estimation, and a new regularity condition requiring a fixed prediction time window $$u_0$$ u 0 is established to consistently estimate the covariance of the estimated parameters. Extensive simulation studies are conducted to assess the properties of the proposed model alongside with an explicit demonstration of the implementation using existing statistical software. The proposed method is applied to a longitudinal Alzheimer’s disease dataset from the National Alzheimer’s Coordinating Center to assess effect of time-dependent cognitive complaint on Alzheimer’s Disease progression.

Mathematical modeling of hypoxia and adenosine to explore tumor escape mechanisms in DC-based immunotherapy

Identifying and controlling tumor escape mechanisms is crucial for improving cancer treatment effectiveness. Experimental studies reveal tumor hypoxia and adenosine as significant contributors to such mechanisms. Hypoxia exacerbates adenosine levels in the tumor microenvironment. Combining inhibition of these factors with dendritic cell (DC)-based immunotherapy promises improved clinical outcomes. However, challenges include understanding dynamics, optimal vaccine dosages, and timing. Mathematical models, including agent-based, diffusion, and ordinary differential equations, address these challenges. Here, we employ these models for the first time to elucidate how hypoxia and adenosine facilitate tumor escape in DC-based immunotherapy. After parameter estimation using experimental data, we optimize vaccination protocols to minimize tumor growth. Sensitivity analysis highlights adenosine’s significant impact on immunotherapy efficacy. Its suppressive role impedes treatment success, but inhibiting adenosine could enhance therapy, as suggested by the model. Our findings shed light on hypoxia and adenosine-mediated tumor escape mechanisms, informing future treatment strategies. Additionally, identifiability analysis confirms accurate parameter determination using experimental data.

The Lomax-Exponentiated Odds Ratio–G Distribution and Its Applications

This paper introduces the Lomax-exponentiated odds ratio–G (L-EOR–G) distribution, a novel framework designed to adeptly navigate the complexities of modern datasets. It blends theoretical rigor with practical application to surpass the limitations of traditional models in capturing complex data attributes such as heavy tails, shaped curves, and multimodality. Through a comprehensive examination of its theoretical foundations and empirical data analysis, this study lays down a systematic theoretical framework by detailing its statistical properties and validates the distribution’s efficacy and robustness in parameter estimation via Monte Carlo simulations. Empirical evidence from real-world datasets further demonstrates the distribution’s superior modeling capabilities, supported by compelling various goodness-of-fit tests. The convergence of theoretical precision and practical utility heralds the L-EOR–G distribution as a groundbreaking advancement in statistical modeling, significantly enhancing precision and adaptability. The new model not only addresses a critical need within statistical modeling but also opens avenues for future research, including the development of more sophisticated estimation methods and the adaptation of the model for various data types, thereby promising to refine statistical analysis and interpretation across a wide array of disciplines.

Improving the Estimation of Structural Parameters of a Mixed Conifer–Broadleaf Forest Using Structural, Textural, and Spectral Metrics Derived from Unmanned Aerial Vehicle Red Green Blue (RGB) Imagery

Forest structural parameters are crucial for assessing ecological functions and forest quality. To improve the accuracy of estimating these parameters, various approaches based on remote sensing platforms have been employed. Although remote sensing yields high prediction accuracy in uniform, even-aged, simply structured forests, it struggles in complex structures, where accurately predicting forest structural parameters remains a significant challenge. Recent advancements in unmanned aerial vehicle (UAV) photogrammetry have opened new avenues for the accurate estimation of forest structural parameters. However, many studies have relied on a limited set of remote sensing metrics, despite the fact that selecting appropriate metrics as powerful explanatory variables and applying diverse models are essential for achieving high estimation accuracy. In this study, high-resolution RGB imagery from DJI Matrice 300 real-time kinematics was utilized to estimate forest structural parameters in a mixed conifer–broadleaf forest at the University of Tokyo Hokkaido Forest (Hokkaido, Japan). Structural and textual metrics were extracted from canopy height models, and spectral metrics were extracted from orthomosaics. Using random forest and multiple linear regression models, we achieved relatively high estimation accuracy for dominant tree height, mean tree diameter at breast height, basal area, mean stand volume, stem density, and broadleaf ratio. Including a large number of explanatory variables proved advantageous in this complex forest, as its structure is influenced by numerous factors. Our results will aid foresters in predicting forest structural parameters using UAV photogrammetry, thereby contributing to sustainable forest management.

Novel estimation of African swine fever transmission parameters within smallholder villages in Lao P.D.R

African Swine Fever (ASF) disease transmission parameters are crucial for making response and control decisions when faced with an outbreak, yet they are poorly quantified for smallholder and village contexts within Southeast Asia. Whilst disease-specific factors − such as latent and infectious periods − should remain reasonably consistent, host, environmental and management factors are likely to affect the rate of disease spread. These differences are investigated using Approximate Bayesian Computation with Sequential Monte-Carlo methods to provide disease parameter estimates in four naïve pig populations in villages of Lao People’s Democratic Republic. The villages represent smallholder pig farmers of the Northern province of Oudomxay and the Southern province of Savannakhet, and the model utilised field mortality data to validate the transmission parameter estimates over the course of multiple model generations. The basic reproductive number between-pigs was estimated to range from 3.08 to 7.80, whilst the latent and infectious periods were consistent with those published in the literature for similar genotypes in the region (4.72 to 6.19 days and 2.63 to 5.50 days, respectively). These findings demonstrate that smallholder village pigs interact similarly to commercial pigs, however the spread of disease may occur slightly slower than in commercial study groups. Furthermore, the findings demonstrated that despite diversity across the study groups, the disease behaved in a consistent manner. This data can be used in disease control programs or for future modelling of ASF in smallholder contexts.

The Extended Bregman Divergence and Parametric Estimation in Continuous Models

The Extended Bregman Divergence and Parametric Estimation in Continuous Models

Untangling the stranglehold through mathematical modelling of Streptococcus equi subspecies equi transmission

Strangles, a disease caused by infection with Streptococccus equi subspecies equi (S. equi), is endemic worldwide and one of the most frequently diagnosed infectious diseases of horses. Recent work has improved our knowledge of key parameters of transmission dynamics, but important knowledge gaps remain. Our aim was to apply mathematical modelling of S. equi transmission dynamics to prioritise future research areas, and add precision to estimates of transmission parameters thereby improving understanding of S. equi epidemiology and quantifying the control effort required. A compartmental deterministic model was constructed. Parameter values were estimated from current literature wherever possible. We assessed the sensitivity of estimates for the basic reproduction number on the population scale to varying assumptions for the unknown or uncertain parameters of: (mean) duration of carriership (1∕γC), relative infectiousness of carriers (f), proportion of infections that result in carriership (p), and (mean) duration of immunity after natural infection (1∕γR). Available incidence and (sero-)prevalence data were compared to model outputs to improve point estimates and ranges for these currently unknown or uncertain transmission-related parameters. The required vaccination coverage of an ideal vaccine to prevent major outbreaks under a range of control scenarios was estimated, and compared available data on existing vaccines. The relative infectiousness of carriers (as compared to acutely ill horses) and the duration of carriership were identified as key knowledge gaps. Deterministic compartmental simulations, combined with seroprevalence data, suggest that 0.05<fˆ<0.5 and that the duration of protective immunity after infection is likely 4–6 years. The presence of carriers alone may suffice to keep S. equi endemic in a population, implying that carriers cannot be ignored in control efforts. Weekly screening of herds for signs of strangles could be sufficient to ensure R < 1, provided all horses are screened for carriership post-infection. In some of worst-case scenarios, vaccination alone would not suffice to prevent major outbreaks from occurring. A stochastic agent-based model was also constructed and validated, and used to simulate a remount depot, to evaluate whether historical incidence data of recurrence of strangles within individuals could be explained without the assumption that one in four horses fail to mount a lasting immune response. These simulations demonstrated that the observed data could have occurred without that assumption.