Beta Mixture Research Articles

Predicting Health Utilities Using Health Administrative Data: Leveraging Survey-linked Health Administrative Data from Ontario, Canada.

The quality-adjusted life year (QALY) is widely used to measure health outcome that combines the length of life and health-related quality of life (HRQoL). To be a reliable QALY measure, HRQoL measurements with a preference-based scoring algorithm need to be converted into health utilities on a scale from zero (dead) to one (perfect health). However, preference-based health utility data are often not available. We address this gap by developing a predictive model for health utilities. To develop a predictive model for health utilities using available demographic and morbidity variables in a health administrative dataset for non-institutionalised populations in Ontario, Canada. The data were obtained from the 2009 to 2010 Canadian Community Health Survey containing Health Utilities Index Mark3 (HUI3), a generic multi-attribute preference-based health utility instrument linked with Ontario health administrative (OHA) data that were collected for administrative or billing purposes for patient encounters with the health care system. We employed four regression models (linear, Tobit, single-part beta mixture, and two-part beta mixture) and a calibration technique to identify the best-fit regression model. Our findings indicate that the two-part beta mixture model is the best-fit for predicting health utilities in the OHA data. The proposed predictive model reflects the original distribution of HUI3 in the population. Our proposed predictive model generates reasonably accurate health utility predictions from OHA data. Our model-based prediction approach is a useful strategy for real-world applications, particularly when preference-based utility data are unavailable.

A novel family of beta mixture models for the differential analysis of DNA methylation data: An application to prostate cancer.

Identifying differentially methylated cytosine-guanine dinucleotide (CpG) sites between benign and tumour samples can assist in understanding disease. However, differential analysis of bounded DNA methylation data often requires data transformation, reducing biological interpretability. To address this, a family of beta mixture models (BMMs) is proposed that (i) objectively infers methylation state thresholds and (ii) identifies differentially methylated CpG sites (DMCs) given untransformed, beta-valued methylation data. The BMMs achieve this through model-based clustering of CpG sites and by employing parameter constraints, facilitating application to different study settings. Inference proceeds via an expectation-maximisation algorithm, with an approximate maximization step providing tractability and computational feasibility. Performance of the BMMs is assessed through thorough simulation studies, and the BMMs are used for differential analyses of DNA methylation data from a prostate cancer study. Intuitive and biologically interpretable methylation state thresholds are inferred and DMCs are identified, including those related to genes such as GSTP1, RASSF1 and RARB, known for their role in prostate cancer development. Gene ontology analysis of the DMCs revealed significant enrichment in cancer-related pathways, demonstrating the utility of BMMs to reveal biologically relevant insights. An R package betaclust facilitates widespread use of BMMs.

Bivariate BMM-based hybrid domain image watermark detector

Robustness, imperceptibility, and watermark capacity are three indispensable and contradictory properties for any image watermarking systems. It is a challenging work to achieve the balance among the three important properties. In this paper, by using the bivariate Beta mixture model (Bivariate BMM), we present a statistical image watermark scheme in nonsubsampled Contourlet transform (NSCT)-fractional order Jacobi Fourier moments (FJFMs) amplitude hybrid domain. The whole watermarking algorithm includes two parts: watermark embedding and detection. NSCT is firstly performed on host image to obtain the frequency subbands, and the NSCT subbands are divided into non overlapping blocks. Then, the significant NSCT domain blocks are selected using local binary patterns (LBP). Meanwhile, for each selected NSCT coefficient block, FJFMs are calculated to obtain the NSCT-FJFMs amplitude. Finally, watermark signals are inserted into the amplitude hybrid domain of NSCT-FJFMs. In order to detect accurately watermark signal, the statistical characteristics of NSCT-FJFMs magnitudes are analyzed in detail. Then, NSCT-FJFMs magnitudes are described statistically by Bivariate BMM, which can simultaneously capture the marginal distribution and strong dependencies of NSCT-FJFMs magnitudes. Also, Bivariate BMM parameters are estimated accurately by the rough-enhanced-Bayes mixture estimation & expectation–maximization (REBMIX&EM) approach. Finally, a statistical watermark detector based on the locally optimum (LO) decision rule and Bivariate BMM is developed in NSCT-FJFMs magnitude hybrid domain. Also, the closed-form expressions on the LO statistic is derived and the receiver operating characteristic (ROC) about our detector is analyzed. Extensive experimental results show the superiority of the proposed image watermark detector over some state-of-the-art statistical watermarking methods.

Probability graph complementation contrastive learning

Graph Neural Network (GNN) has achieved remarkable progress in the field of graph representation learning. The most prominent characteristic, propagating features along the edges, degrades its performance in most heterophilic graphs. Certain researches make attempts to construct KNN graph to improve the graph homophily. However, there is no prior knowledge to choose proper K and they may suffer from the problem of Inconsistent Similarity Distribution (ISD). To accommodate this issue, we propose Probability Graph Complementation Contrastive Learning (PGCCL) which adaptively constructs the complementation graph. We employ Beta Mixture Model (BMM) to distinguish intra-class similarity and inter-class similarity. Based on the posterior probability, we construct Probability Complementation Graphs to form contrastive views. The contrastive learning prompts the model to preserve complementary information for each node from different views. By combining original graph embedding and complementary graph embedding, the final embedding is able to capture rich semantics in the finetuning stage. At last, comprehensive experimental results on 20 datasets including homophilic and heterophilic graphs firmly verify the effectiveness of our algorithm as well as the quality of probability complementation graph compared with other state-of-the-art methods.

Bayesian Estimation of Inverted Beta Mixture Models With Extended Stochastic Variational Inference for Positive Vector Classification.

The finite inverted beta mixture model (IBMM) has been proven to be efficient in modeling positive vectors. Under the traditional variational inference framework, the critical challenge in Bayesian estimation of the IBMM is that the computational cost of performing inference with large datasets is prohibitively expensive, which often limits the use of Bayesian approaches to small datasets. An efficient alternative provided by the recently proposed stochastic variational inference (SVI) framework allows for efficient inference on large datasets. Nevertheless, when using the SVI framework to address the non-Gaussian statistical models, the evidence lower bound (ELBO) cannot be explicitly calculated due to the intractable moment computation. Therefore, the algorithm under the SVI framework cannot directly use stochastic optimization to optimize the ELBO, and an analytically tractable solution cannot be derived. To address this problem, we propose an extended version of the SVI framework with more flexibility, namely, the extended SVI (ESVI) framework. This framework can be used in many non-Gaussian statistical models. First, some approximation strategies are applied to further lower the ELBO to avoid intractable moment calculations. Then, stochastic optimization with noisy natural gradients is used to optimize the lower bound. The excellent performance and effectiveness of the proposed method are verified in real data evaluation.

Personnel dosimetry for the mixed beta (Kr-85)/photon (gamma, X-ray) beams using Harshaw 8814 TL dosimeters: Performance Test Applicability Study

Performance of Harshaw 8814 TL dosimeters was evaluated in the mixture of low energy beta (Kr-85, average energy 251 keV) and X-ray/gamma photon beams, when many doses obtained using existing dose calculation algorithms show a high bias. 264 TLDs were exposed at the Radiological and Environmental Sciences Laboratory to the pure Kr-85 energy beam as well as to different combinations of beta and photon beams of various energies (including Cs-137 and Co-60 gamma sources and M50, M200 and H300 X-ray beams). The TLDs were irradiated to different beta-to-photon dose ratios; 22 different combinations of radiation energy beams and beam dose ratios were tested. After exposure, the TLDs were readout at Brookhaven National Laboratory facility using a Harshaw 8800 reader. The doses were determined with the Harshaw Combined DOELAP dose calculation algorithm and compared to the known values. The results showed a significant deviation from the corresponding nominal values for some of the tested TLDs exposed to pure Kr-85 or a combination of Kr-85 and photon beam mixed energies. The different ratios of TL signals from chips 1–4 of the 8814 TL cards were analyzed, and an algorithm was proposed to identify the radiation beam energy to which each individual TLD was irradiated to. Finally, correction coefficients were developed which improves the performance of TL dosimetry dose calculation algorithm in the cases of exposure to a low energy beta beam and to a mixture of such beam with photons.

Development of algorithms to estimate the EQ-5D-5L from the FACT-L in patients with lung cancer: a mapping study.

This study aimed to develop a mapping algorithm to evaluate the EQ-5D-5L according to the FACT-L when the EQ-5D-5L is not available. EQ-5D-5L and FACT-L data were collected from patients with lung cancer in Departments of Thoracic Surgery, Medical Oncology, Radiation Oncology, Sichuan Cancer Hospital. We used the ordinary least squares model (OLS), Tobit model (Tobit), two-part model (TPM), beta mixture regression (BM), and censored least absolute deviation model (CLAD) to map the results of the FACT-L according to EQ-5D-5L scores. To establish these models, the total score, dimension scores, squared items, and interaction items were introduced. Performance metrics including Adjusted R2, root mean square error (RMSE), and mean absolute error (MAE) were used to select the optimized model. The model with the best mapping performance was the BM model (BETAMIX4) with the PWB (physical well-being) dimension, FWB (functional well-being) dimension, the squared term of the PWB dimension, and the squared term of the FWB dimension as covariates. The final prediction metrics were Adjusted R2 = 0.695, RMSE = 0.206, and MAE = 0.109. Fivefold cross-validation (CV) results also demonstrated that the BM model had the best mapping power. This study developed an optimized mapping algorithm to predict the utility index from the FACT-L to the EQ-5D-5L, which provides an effective alternative reference for EQ-5D-5L estimation when the preference-based health utility values were unavailable.

ScDMV: a zero-one inflated beta mixture model for DNA methylation variability with scBS-seq data.

The utilization of single-cell bisulfite sequencing (scBS-seq) methods allows for precise analysis of DNA methylation patterns at the individual cell level, enabling the identification of rare populations, revealing cell-specific epigenetic changes, and improving differential methylation analysis. Nonetheless, the presence of sparse data and an overabundance of zeros and ones, attributed to limited sequencing depth and coverage, frequently results in reduced precision accuracy during the process of differential methylation detection using scBS-seq. Consequently, there is a pressing demand for an innovative differential methylation analysis approach that effectively tackles these data characteristics and enhances recognition accuracy. We propose a novel beta mixture approach called scDMV for analyzing methylation differences in single-cell bisulfite sequencing data, which effectively handles excess zeros and ones and accommodates low-input sequencing. Our extensive simulation studies demonstrate that the scDMV approach outperforms several alternative methods in terms of sensitivity, precision, and controlling the false positive rate. Moreover, in real data applications, we observe that scDMV exhibits higher precision and sensitivity in identifying differentially methylated regions, even with low-input samples. In addition, scDMV reveals important information for GO enrichment analysis with single-cell whole-genome sequencing data that are often overlooked by other methods. The scDMV method, along with a comprehensive tutorial, can be accessed as an R package on the following GitHub repository: https://github.com/PLX-m/scDMV.

Mixed-type defect pattern recognition in noisy labeled wafer bin maps

In semiconductor manufacturing, classification of defect patterns in wafer bin maps (WBMs) helps engineers detect process failures and identify their causes. In recent studies on WBMs, convolutional neural networks (CNNs) have demonstrated effective classification performance based on their high expressive power. However, previous studies have implicitly assumed that the labels of WBMs used for training CNNs are correct, even though labels are often incorrect. When trained on mislabeled data, CNNs with standard cross-entropy loss can easily overfit mislabeled samples, leading to poor generalization for testing data. To overcome this issue, we propose a novel training algorithm called sample bootstrapping. Sample bootstrapping identifies which samples have clean or noisy labels by using a two-component beta mixture model, and measures the uncertainty of each identified label. Then, samples with low uncertainty of their estimated labels are selected to build mini-batches via weighted random sampling. Finally, CNNs are trained on the selected mini-batches with dynamic bootstrapping loss. In this manner, we can correct only the samples that are highly likely to have noisy labels and prevent the risk of false correction of correctly labeled samples. Experiments on simulated and real datasets demonstrate the effectiveness of the proposed method.

Comparison of combination methods to create calibrated ensemble forecasts for seasonal influenza in the U.S.

The characteristics of influenza seasons vary substantially from year to year, posing challenges for public health preparation and response. Influenza forecasting is used to inform seasonal outbreak response, which can in turn potentially reduce the impact of an epidemic. The United States Centers for Disease Control and Prevention, in collaboration with external researchers, has run an annual prospective influenza forecasting exercise, known as the FluSight challenge. Uniting theoretical results from the forecasting literature with domain-specific forecasts from influenza outbreaks, we applied parametric forecast combination methods that simultaneously optimize model weights and calibrate the ensemble via a beta transformation and made adjustments to the methods to reduce their complexity. We used the beta-transformed linear pool, the finite beta mixture model, and their equal weight adaptations to produce ensemble forecasts retrospectively for the 2016/2017, 2017/2018, and 2018/2019 influenza seasons in the U.S. We compared their performance to methods that were used in the FluSight challenge to produce the FluSight Network ensemble, namely the equally weighted linear pool and the linear pool. Ensemble forecasts produced from methods with a beta transformation were shown to outperform those from the equally weighted linear pool and the linear pool for all week-ahead targets across in the test seasons based on average log scores. We observed improvements in overall accuracy despite the beta-transformed linear pool or beta mixture methods' modest under-prediction across all targets and seasons. Combination techniques that explicitly adjust for known calibration issues in linear pooling should be considered to improve probabilistic scores in outbreak settings.

Obtaining SF-6D utilities from FACT-H&N in thyroid carcinoma patients: development and results from a mapping study.

There is limited evidence for mapping clinical tools to preference-based generic tools in the Chinese thyroid cancer patient population. The current study aims to map the FACT-H&N (Functional Assessment of Cancer Therapy-Head and Neck Cancer) to the SF-6D (Short Form Six-Dimension), which will inform future cost-utility analyses related to thyroid cancer treatment. A total of 1050 participants who completed the FACT-H&N and SF-6D questionnaires were included in the analysis. Four methods of direct and indirect mapping were estimated: OLS regression, Tobit regression, ordered probit regression, and beta mixture regression. We evaluated the predictive performance in terms of root mean square error (RMSE), mean absolute error (MAE), concordance correlation coefficient (CCC), Akaike information criterion (AIC) and Bayesian information criterion (BIC) and the correlation between the observed and predicted SF-6D scores. The mean value of SF-6D was 0.690 (SD = 0.128). The RMSE values for the fivefold cross-validation as well as the 30% random sample validation for multiple models in this study were 0.0833-0.0909, MAE values were 0.0676-0.0782, and CCC values were 0.6940-0.7161. SF-6D utility scores were best predicted by a regression model consisting of the total score of each dimension of the FACT-H&N, the square of the total score of each dimension, and covariates including age and gender. We proposed to use direct mapping (OLS regression) and indirect mapping (ordered probit regression) to establish a mapping model of FACT-H&N to SF-6D. The mean SF-6D and cumulative distribution functions simulated from the recommended mapping algorithm generally matched the observed ones. In the absence of preference-based quality of life tools, obtaining the health status utility of thyroid cancer patients from directly mapped OLS regression and indirectly mapped ordered probit regression is an effective alternative.

Small area estimation of inequality measures using mixtures of Beta

Abstract Economic inequalities referring to specific regions are crucial in deepening spatial heterogeneity. Income surveys are generally planned to produce reliable estimates at countries or macroregion levels, thus we implement a small area model for a set of inequality measures (Gini, Relative Theil, and Atkinson indexes) to obtain reliable microregion estimates. Considering that inequality estimators are unit-interval defined with skewed and heavy-tailed distributions, we propose a Bayesian hierarchical model at the area level involving a Beta mixture. An application on EU-SILC data is carried out and a design-based simulation is performed. Our model outperforms in terms of bias, coverage, and error the standard Beta regression model. Moreover, we extend the analysis of inequality estimators by deriving their approximate variance functions.

Enabling QALY estimation in mental health trials and care settings: mapping from the PHQ-9 and GAD-7 to the ReQoL-UI or EQ-5D-5L using mixture models

PurposePatient-reported outcome measures (PROMs) are commonly collected in trials and some care settings, but preference-based PROMs required for economic evaluation are often missing. For these situations, mapping models are needed to predict preference-based (aka utility) scores. Our objective is to develop a series of mapping models to predict preference-based scores from two mental health PROMs: Patient Health Questionnaire-9 (PHQ-9; depression) and Generalised Anxiety Questionnaire-7 (GAD-7; anxiety). We focus on preference-based scores for the more physical-health-focussed EQ-5D (five-level England and US value set, and three-level UK cross-walk) and more mental-health-focussed Recovering Quality-of-Life Utility Index (ReQoL-UI).MethodsWe used trial data from the Improving Access to Psychological Therapies (IAPT) mental health services (now called NHS Talking Therapies), England, with a focus on people with depression and/or anxiety caseness. We estimated adjusted limited dependent variable or beta mixture models (ALDVMMs or Betamix, respectively) using GAD-7, PHQ-9, age, and sex as covariates. We followed ISPOR mapping guidance, including assessing model fit using statistical and graphical techniques.ResultsOver six data collection time-points between baseline and 12-months, 1340 observed values (N ≤ 353) were available for analysis. The best fitting ALDVMMs had 4-components with covariates of PHQ-9, GAD-7, sex, and age; age was not a probability variable for the final ReQoL-UI mapping model. Betamix had practical benefits over ALDVMMs only when mapping to the US value set.ConclusionOur mapping functions can predict EQ-5D-5L or ReQoL-UI related utility scores for QALY estimation as a function of variables routinely collected within mental health services or trials, such as the PHQ-9 and/or GAD-7.

DNA Methylation in Risk for Cognitive Impairment and Type 2 Diabetes in a Mexican American Cohort

AbstractBackgroundAlzheimer’s disease (AD) and type 2 diabetes (T2D) are among the leading causes of mortality among the growing aging Mexican American population (≥ 65 years old) in the US (Vega et al. 2017). In comparison to their non‐Hispanic white counterparts who most likely develop inflammation associated AD, aging Mexican Americans have an earlier onset of AD and metabolism related predisposition for AD (O'Bryant et al. 2010). Mild cognitive impairment (MCI) is a phenotype that often leads to AD and is also prevalent in this cohort (O'Bryant et al. 2013). The risks for AD, MCI and T2D are multifactorial, involving a form of epigenetic regulation called methylation where a methyl group is added to the cytosine base in DNA (Shao et al. 2017; Juvinao‐Quintero et al. 2019). We aim to elucidate an epigenetic association between cognitive impairment (identified here as AD and MCI), and T2D that is unique to the Mexican American population.MethodA total of 551 aging participants from the Texas Alzheimer’s Research and Care Consortium (TARCC) consisting of 252 non‐Hispanic white individuals and 299 Mexican Americans were selected, after quality control. First, participants were stratified into groups of individuals diagnosed with cognitive impairment (CI) alone and controls without CI within each ethnic group. Secondly, this cohort was stratified into individuals with T2D alone and controls without T2D. Thirdly, participants were stratified into those with both CI and T2D versus normal healthy controls. Lastly, any differential methylation associated with each ethnic group will be compared and contrasted. Peripheral blood drawn from participants was used to obtain individual methylation profiles using the Illumina Infinium MethylationEPIC chip array. Differential methylation was assessed using the Chip Analysis Methylation Pipeline (ChAMP), limma and cate packages in R. The Beta MIxture Quantile dilation (BMIQ) method was used for data normalization.ResultResults will be analyzed using gene set enrichment and pathway analysis tools.ConclusionIdentifying possible methylation sites associated with CI and T2D could contribute towards developing ethnicity‐specific biomarkers for Mexican Americans.

Mapping study of papillary thyroid carcinoma in China: Predicting EQ-5D-5L utility values from FACT-H&N.

To develop a mapping algorithm that can be used to predict EQ-5D-5L health utility scores from FACT-H&N and obtain health utility parameters for Chinese patients with papillary thyroid carcinoma (PTC), which can be used for cost-utility analysis in health economic. A total of 1,050 patients with PTC from a tertiary hospital in China were included, and they completed FACT-H&N and EQ-5D-5L. Four mapping algorithms of direct mapping functions were used to derive the models: Ordinary least squares (OLS), Tobit model (Tobit), Two-part model (TPM), and Beta mixture regression model (Beta). The goodness-of-fit of models was assessed by the mean absolute error (MAE), root mean square error (RMSE), Akaike information criteria (AIC), Bayesian information criteria (BIC), and absolute error (AE). A fivefold cross-validation method was used to test the stability of the models. The mean utility value of the EQ-5D-5L was 0.870 ± 0.094. The mean EQ-VAS score was 76.5 ± 13.0. The Beta mixture regression model mapping FACT-H&N to EQ-5D-5L achieved the best performance [fivefold cross-validation MAE = 0.04612, RMSE = 0.06829, AIC = -2480.538, BIC = -2381.137, AE > 0.05 (%) = 32.48, AE > 0.1 (%) = 8.95]. The independent variables in this model were Physical Well-Being (PWB), Emotional Well-Being (EWB), Head & Neck Cancer Subscale (HNCS) scores and its square term and interaction term scores. This study calculated the health utility score of Chinese patients with PTC. The reported algorithms can be used to map the FACT-H&N into the EQ-5D-5L, which can be applied in the cost-utility related study of patients with PTC.

A Stochastic and Bayesian Inference Toolchain for Uncertainty and Risk Quantification of Rare Autoignition Events in Dry Low-Emission Premixers

Abstract Quantification of aleatoric uncertainties due to the inherent variabilities in operating conditions and fuel composition is essential for designing and improving premixers in dry low-emissions (DLE) combustion systems. Advanced stochastic simulation tools require a large number of evaluations in order to perform this type of uncertainty quantification (UQ) analysis. This task is computationally prohibitive using high-fidelity computational fluid dynamic (CFD) approaches such as large eddy simulation (LES). In this paper, we describe a novel and computationally efficient toolchain for stochastic modeling using minimal input from LES, to perform uncertainty and risk quantification of a DLE system. More specially, high-fidelity LES, chemical reactor network (CRN) model, beta mixture model, Bayesian inference and sequential Monte Carlo (SMC) are integrated into the toolchain. The methodology is applied to a practical premixer of low-emission combustion system with dimethyl ether (DME)/methane–air mixtures to simulate auto-ignition events at different engine conditions. First, the benchmark premixer is simulated using a set of LESs for a methane/air mixture at elevated pressure and temperature conditions. A partitioning approach is employed to generate a set of deterministic chemical reactor network (CRN) models from LES results. These CRN models are then solved at the volume-average conditions and validated by LES results. A mixture modeling approach using the expectation-method of moment (E-MM) is carried out to generate a set of beta mixture models and characterize uncertainties for LES-predicted temperature distributions. These beta mixture models and a normal distribution for DME volume fraction are used to simulate a set of stochastic CRN models. The Bayesian inference approach through SMC method is then implemented on the results of temperature distributions from stochastic CRN models to simulate the probability of auto-ignition in the benchmark premixer. The results present a very satisfactory performance for the stochastic toolchain to compute the auto-ignition propensity for a few events with a particular combination of inlet temperature and DME volume fraction. Characterization of these rare events is computationally prohibitive in the conventional deterministic methods such as high-fidelity LES.

Health-related quality of life in patients with multiple system atrophy using the EQ-5D-5L.

BackgroundMultiple system atrophy (MSA) is an incurable neurodegenerative disease. We aimed to investigate the health‐related quality of life (HRQoL) and the determinants of HRQoL in patients with MSA.MethodsThe five‐level EuroQol five‐dimensional questionnaire (EQ‐5D‐5L) was used to evaluate patients’ HRQoL. The results of HRQoL were indicated by the EQ‐5D‐5L index values and visual analog scale (EQ VAS) scores. Specific scales were used to measure disease severity, cognition, frontal lobe function, anxiety, depression, fatigue, and sleep disorders. The beta mixture model and the linear regression model were used to explore the determinants of HRQoL in patients with MSA.ResultsA total of 205 patients with cerebellar variants (MSA‐C; 53.9%) and 175 patients with parkinsonian variants (MSA‐P; 46.1%) were included in this cross‐sectional study. The mean values of the EQ‐5D‐5L index values and EQ VAS scores were .558 and 59.5, respectively. Problem with mobility was the problem reported by the highest proportion of patients (92.1%), followed by problems with usual activities (88.7%), self‐care (81.3%), anxiety/depression (72.1%), and pain/discomfort (53.9%). The determinants of the lower EQ‐5D‐5L index values in patients with MSA were greater disease severity, fatigue, Parkinson's disease‐related sleep problems (PD‐SP), depressive mood, and anxious mood. Greater disease severity, fatigue, and depressive mood were associated with lower EQ VAS scores.ConclusionThe problem reported most frequently by Chinese individuals with MSA was mobility. In addition to the greater disease severity of MSA, fatigue, PD‐SP, depression, and anxiety were determinants of poor HRQoL.

Beta Mixture Model for the Uncertainties in Robotic Haptic Object Identification

Robotic haptic object identification is the process to identify objects out of a given object set using a robotic hand equipped with tactile and finger-joint displacement sensors. When taking measurements by grasping the object, the uncertainties in the pose of the object relative to the hand will adversely affect the identification accuracy. Each tactile sensor measures contact in its locality, thus, a change in object contact locations relative to the robotic grasping hand significantly affects the tactile measurements. In object identification, statistical properties of the uncertainties in the collected measurements are generally obtained <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">a priori</i> , allowing the probabilities of an object to be estimated for improved accuracy. The problem of object pose uncertainty typical in robotic grasping results in multiple peaks in the probability distribution of the resulting tactile measurements. The peaks are associated with whether or not the (locality of) tactile sensor on the robotic hand is in contact with the object due to the variations in object pose. As such, in this article, a Beta mixture model allowing multiple peaks in the distribution is proposed to represent this object pose uncertainty (relative to the robotic hand) in place of the conventional Gaussian model used in the literature. The method was experimentally validated and demonstrated to be effective in capturing the uncertainties and improving the accuracy of the haptic object identification.

Bayesian flexible beta regression model with functional covariate

Standard parametric regression models are unsuitable when the aim is to predict a bounded continuous response, such as a proportion/percentage or a rate. A possible solution is the flexible beta regression model which is based on a special mixture of betas designed to cope with (though not limited to) bimodality, heavy tails, and outlying observations. This work introduces such a model in the case of a functional covariate, motivated by a spectrometric analysis on milk specimens. Estimation issues are dealt with through a combination of standard basis expansion and Markov chains Monte Carlo techniques. Specifically, the selection of the most significant coefficients of the expansion is done through Bayesian variable selection methods that take advantage of shrinkage priors. The effectiveness of the proposal is illustrated with simulation studies and the application on spectrometric data.

A Non-parametric Bayesian Learning Model Using Accelerated Variational Inference on Multivariate Beta Mixture Models for Medical Applications

Clustering as an exploratory technique has been a promising approach for performing data analysis. In this paper, we propose a non-parametric Bayesian inference to address clustering problem. This approach is based on infinite multivariate Beta mixture models constructed through the framework of Dirichlet process. We apply an accelerated variational method to learn the model. The motivation behind proposing this technique is that Dirichlet process mixture models are capable to fit the data where the number of components is unknown. For large-scale data, this approach is computationally expensive. We overcome this problem with the help of accelerated Dirichlet process mixture models. Moreover, the truncation is managed using kd-trees. The performance of the model is validated on real medical applications and compared to three other similar alternatives. The results show the outperformance of our proposed framework.