Knowing What to Ask: A Bayesian Active Learning Approach to the Surveying Problem

The effect of missing data on phylogenetic methods is a potentially important issue in our attempts to reconstruct the Tree of Life. If missing data are truly problematic, then it may be unwise to include species in an analysis that lack data for some characters (incomplete taxa) or to include characters that lack data for some species. Given the difficulty of obtaining data from all characters for all taxa (e.g., fossils), missing data might seriously impede efforts to reconstruct a comprehensive phylogeny that includes all species. Fortunately, recent simulations and empirical analyses suggest that missing data cells are not themselves problematic, and that in- complete taxa can be accurately placed as long as the overall number of characters in the analysis is large. How- ever, these studies have so far only been conducted on parsimony, likelihood, and neighbor-joining methods. Although Bayesian phylogenetic methods have become widely used in recent years, the effects of missing data on Bayesian analysis have not been adequately studied. Here, we conduct simulations to test whether Bayesian analyses can accurately place incomplete taxa despite extensive missing data. In agreement with previous studies of other methods, we find that Bayesian analyses can accurately reconstruct the position of highly incomplete taxa (i.e., 95% missing data), as long as the overall number of characters in the analysis is large. These results suggest that highly incomplete taxa can be safely included in many Bayesian phylogenetic analyses. The impact of missing data is a potentially im- portant issue in phylogenetic analyses, particularly if the goal is to reconstruct a comprehensive Tree of Life that includes both fossil and living taxa. Missing data are often encountered when combining data from two or more different genes, when some of the taxa have sequence data available for one gene but not the other. If the taxa lacking data for a gene are included in the combined analysis, then the characters associated with this gene are typically coded as missing or unknown (often denoted with a ?). Similarly, missing data are often encountered in analyses that include fossil taxa, when certain taxa must be scored as unknown for certain characters because the relevant features have not been adequately preserved. Concerns about missing data may often deter- mine what characters and taxa will be included in an analysis (Wiens, 2006), even if this is not always stated explicitly by researchers. For example, if missing data are considered to be problematic, then one should only include species that have complete data for all characters or else only include characters that have complete data for all species. Thus, one may have to reduce the number of taxa or characters in an

Active learning approaches require a lot of time investment in student activities and engagement during the class period, which often leads to incomplete coverage of the course syllabus. Furthermore, it requires significant amount of time for the instructor to design and implement active learning strategies. All these shortcomings are often cited as the common hindrance in adopting student active learning. To address these widely recognized inhibiting factors, we recently implemented a new student active learning approach. This approach is based on the students’ presentation and hence termed as student presentation-based active learning (SPAL). Under this approach, students are given a reading assignment to prepare a PowerPoint presentation on well-defined conceptual topics, questions, or chapter modules. Reading assignments on a topic are administered 1–2 weeks before covering them in the class. This allows reasonable time for the self-comprehension of the suggested material for presentation preparation. Students were expected to rehearse the presentation and be prepared to complete it in the suggested time duration. During each lecture, one group of student would present the assigned topic to the class, and their presentations were graded according to the rubric focusing on the coverage of suggested topics, quality of presentation, and after presentation discussions. Peers and instructor provided feedback about the students’ presentation and unclear concepts. To understand the efficacy of this approach, this approach was implemented in Mechanical Engineering Senior Capstone Project I course in Fall 2016. Here the responses and insights garnered from this practice were presented, and discussion on the advantages and challenges associated with the adaptation of this approach in teaching engineering courses as compared to lecture based classroom education system. This paper can provide useful insights for instructors considering this approach or similar student active teaching approach in their courses.

Recent developments in the field of Information Technology call for a specific reconsideration of the traditional educational methods and facilitate development of new teaching and learning methods and tools, opening it for a wider scope of concerned persons. Active teaching and learning methods as part of educational methodology are widely used at the universities across the European community, putting the larger responsibility of learning at the hands of the learners themselves. One of the major challenges is effective addressing the learning needs of the new generation of students. Postgraduate education in in the field of logistics and supply chain management (LSCM) must keep it in mind in order to promote a suitable answer to the LSCM labor market needs. This is accomplished well by a master programme in LSCM with a European dimension emerged from the synergy of international cooperation between Academia and the professional bodies. This paper shares the empirical study on implementation of Master study programme of Logistics and Supply Chain Management in the Riga Technical University and extensive use of active and innovative learning approach.

Active and machine learning-based approaches to rapidly enhance microbial chemical production

Kepler has revolutionized our understanding of both exoplanets and their host stars. Asteroseismology is a valuable tool in the characterization of stars and Kepler is an excellent observing facility to perform asteroseismology. Here we select a sample of 35 Kepler solar-type stars which host transiting exoplanets (or planet candidates) with detected solar-like oscillations. Using available Kepler short cadence data up to Quarter 16 we create power spectra optimized for asteroseismology of solar-type stars. We identify modes of oscillation and estimate mode frequencies by 'peak bagging' using a Bayesian Markov Chain Monte Carlo framework. In addition, we expand the methodology of quality assurance using a Bayesian unsupervised machine learning approach. We report the measured frequencies of the modes of oscillation for all 35 stars and frequency ratios commonly used in detailed asteroseismic modelling. Due to the high correlations associated with frequency ratios we report the covariance matrix of all frequencies measured and frequency ratios calculated. These frequencies, frequency ratios, and covariance matrices can be used to obtain tight constraint on the fundamental parameters of these planet-hosting stars.

Missing data are common in clinical trials and could lead to biased estimation of treatment effects. The National Research Council (NRC) report suggests that sensitivity analysis on missing data mechanism should be a mandatory component of the primary reporting of findings from clinical trials, and regulatory agencies are requesting more thorough sensitivity analyses from sponsors. However, recent literature research showed that missing data were almost always inadequately handled. This is partially due to the lack of standard software packages and straightforward implementation platform. With recent availability of flexible Bayesian software packages such as WinBUGS, SAS Proc MCMC, and Stan, it is relatively simple to develop Bayesian methods to address complex missing data problems while incorporating the uncertainty. In this article, we present a case study from the DIA Bayesian Scientific Working Group (BSWG) on Bayesian approaches for missing data analysis. We illustrate how to use Bayesian approaches to fit a few commonly used frequentist missing data models. The properties, advantage, and flexibility of the Bayesian analysis methods will be discussed using a case study based on a schizophrenia clinical trial. Supplementary materials for this article are available online.

As software continues to insinuate itself into nearly every aspect of our life, the quality of software has been an extremely important issue. Software Quality Assurance (SQA) is a process that ensures the development of high-quality software. It concerns the important problem of maintaining, monitoring, and developing quality software. Accurate detection of fault prone components in software projects is one of the most commonly practiced techniques that offer the path to high quality products without excessive assurance expenditures. This type of quality modeling requires the availability of software modules with known fault content developed in similar environment. However, collection of fault data at module level, particularly in new projects, is expensive and time-consuming. Semi-supervised learning and active learning offer solutions to this problem for learning from limited labeled data by utilizing inexpensive unlabeled data.;In this dissertation, we investigate semi-supervised learning and active learning approaches in the software fault prediction problem. The role of base learner in semi-supervised learning is discussed using several state-of-the-art supervised learners. Our results showed that semi-supervised learning with appropriate base learner leads to better performance in fault proneness prediction compared to supervised learning. In addition, incorporating pre-processing technique prior to semi-supervised learning provides a promising direction to further improving the prediction performance. Active learning, sharing the similar idea as semi-supervised learning in utilizing unlabeled data, requires human efforts for labeling fault proneness in its learning process. Empirical results showed that active learning supplemented by dimensionality reduction technique performs better than the supervised learning on release-based data sets.

Statistics comes in two main flavors: frequentist and Bayesian. For historical and technical reasons, frequentist statistics have traditionally dominated empirical data analysis, and certainly remain prevalent in empirical software engineering. This situation is unfortunate because frequentist statistics suffer from a number of shortcomings---such as lack of flexibility and results that are unintuitive and hard to interpret---that curtail their effectiveness when dealing with the heterogeneous data that is increasingly available for empirical analysis of software engineering practice. In this paper, we pinpoint these shortcomings, and present Bayesian data analysis techniques that provide tangible benefits---as they can provide clearer results that are simultaneously robust and nuanced. After a short, high-level introduction to the basic tools of Bayesian statistics, we present the reanalysis of two empirical studies on the effectiveness of automatically generated tests and the performance of programming languages. By contrasting the original frequentist analyses with our new Bayesian analyses, we demonstrate the concrete advantages of the latter. To conclude we advocate a more prominent role for Bayesian statistical techniques in empirical software engineering research and practice.

Background Accurate estimation of pathogen transmission and importation in healthcare settings typically rely on surveillance cultures, which are costly and not routinely implemented. Clinical cultures, ordered based on symptoms or suspicion are subject to differential test sensitivity, partially capturing asymptomatic colonization, limiting their utility for transmission modeling. We evaluated whether BayesFlow, a Bayesian deep learning approach, can estimate key transmission parameters using clinical culture data without active surveillance. Methods We simulated pathogen spread in a hospital ward where individuals were either susceptible or colonized. Test rates and culture sensitivity varied by colonization status. We compared two data structures: (1) bed-level tracking of test results per bed over time, and (2) agent-level tracking of individuals' admission and testing. BayesFlow was trained on 10,000 simulations and applied to a dataset with known parameters. We estimated transmission, importation, and contact rate parameters and compared simulated transmission dynamics from true and estimated parameters. Results In the Bed-Level structure, BayesFlow substantially overestimated the transmission probability (0.15 vs. true value 0.05) (Table 1), resulting in higher downstream colonizations (Figure 1) while the Fixed Agent structure yielded more accurate estimates across all parameters (Table 2). Simulations based on estimated parameters in this structure produced similar distributions of R₀ and proportion colonized to those generated with true parameters (Figure 1). Conclusion BayesFlow shows promise to estimate transmission and importation parameters using only clinical culture data, particularly when using agent-level representations of hospital dynamics, allowing healthcare epidemiologists to infer key drivers of pathogen spread in the absence of surveillance testing. Accurate estimates of transmission dynamics can inform targeted infection prevention interventions, such as contact precautions or cohorting strategies. Next steps include applying this framework to real-world ICU data to assess its utility for monitoring healthcare-associated infection risk and guiding infection control efforts across diverse hospital settings. Disclosures Karim Khader, PhD, bioMerieux: Grant/Research Support|Merck: Grant/Research Support

Long-lasting and difficult-to-treat, hypertension frequently leads to serious and life-threatening diseases. As a result, early risk assessment and prevention of hypertension are crucial. The majority of research currently available ignore the preprocessing analysis of real medical data, particularly the analysis of missing values, in favor of using clean data to increase the performance of hypertension prediction. Thus, in this study, real but incomplete data were subjected to preprocessing analysis including missing value analysis and feature divergence analysis, and then a Bayesian optimization technique was employed to find the optimal random forest model. Experimental results showed that proper missing value strategy (i.e., MissForest) can slightly enhance the data quality and produce slightly better predictive performance (from 0.001% to 0.069%) even the missing rate is less than 1%. Additionally, compared to using the original features, removing some features with little divergence can lower the dimensionality and even marginally enhance performance by 0.161% in terms of median AUC across 50 runs. Furthermore, the optimal random forest can demonstrate better hypertension discrimination in real medical data. In our case, the optimal random forest can improve the performance of the non-optimized forest by up to 3.51%.

This chapter focuses on measurement error in covariates in regression analyses in which the aim is to estimate the association between one or more covariates and an outcome, adjusting for confounding. Error in covariate measurements, if ignored, results in biased estimates of parameters representing the associations of interest. Studies with variables measured with error can be considered as studies in which the true variable is missing, for either some or all study participants. We make the link between measurement error and missing data and describe methods for correcting for bias due to covariate measurement error with reference to this link, including regression calibration (conditional mean imputation), maximum likelihood and Bayesian methods, and multiple imputation. The methods are illustrated using data from the Third National Health and Nutrition Examination Survey (NHANES III) to investigate the association between the error-prone covariate systolic blood pressure and the hazard of death due to cardiovascular disease, adjusted for several other variables including those subject to missing data. We use multiple imputation and Bayesian approaches that can address both measurement error and missing data simultaneously. Example data and R code are provided in online materials.

Bayesian networks can be used to identify possible causal relationships between variables based on their conditional dependencies and independencies, which can be particularly useful in complex biological scenarios with many measured variables. Here we propose two improvements to an existing method for Bayesian network analysis, designed to increase the power to detect potential causal relationships between variables (including potentially a mixture of both discrete and continuous variables). Our first improvement relates to the treatment of missing data. When there is missing data, the standard approach is to remove every individual with any missing data before performing analysis. This can be wasteful and undesirable when there are many individuals with missing data, perhaps with only one or a few variables missing. This motivates the use of imputation. We present a new imputation method that uses a version of nearest neighbour imputation, whereby missing data from one individual is replaced with data from another individual, their nearest neighbour. For each individual with missing data, the subsets of variables to be used to select the nearest neighbour are chosen by sampling without replacement the complete data and estimating a best fit Bayesian network. We show that this approach leads to marked improvements in the recall and precision of directed edges in the final network identified, and we illustrate the approach through application to data from a recent study investigating the causal relationship between methylation and gene expression in early inflammatory arthritis patients. We also describe a second improvement in the form of a pseudo-Bayesian approach for upweighting certain network edges, which can be useful when there is prior evidence concerning their directions.

Public health programs are increasingly offering health policy content to meet workforce demand for public health professionals with health policy analysis and communication skills. Undergraduate public health education in health policy and related curriculum is lacking. Existing work describes service- and experiential learning approaches, which may not be feasible in large courses or public health programs with limited administrative or financial resources. This article describes health policy curriculum, learning objectives, and course activities (i.e., individual writing assignment, mock policy debate, and policy analysis exercise) that integrate an active learning approach and use simulation to develop policy analysis and communication skills in an undergraduate classroom setting. An active learning approach can increase teaching effectiveness, critical thinking, peer interaction, and student engagement to cultivate these applied skills. The described curriculum has been used in various modalities (in-person, hybrid, and remote) with positive reception in a public research university with a high percentage of first-generation college students. Student course evaluations and an independent assessment of the policy analysis exercise reflect high student interest, engagement, problem solving, satisfaction, perceived benefits, and application of policy analysis and communication skills, providing support for teaching health policy curriculum to undergraduate students using an active learning approach. The discussion includes strategies to address resistance by public health instructors and students to using active learning techniques.

Active learning can serve as a platform for addressing sustainability, ethics, and social entrepreneurship by engaging students in experiential activities with external organisations for initiatives with a social mission. A variety of active learning approaches provide an opportunity to promote collaboration between students and external organisations in pursuit of a socially beneficial goal. This paper highlights different active learning approaches that emphasise sustainability, ethics, and social entrepreneurship in undergraduate and graduate management education. These approaches include experiential learning methods, such as undergraduate case competitions and global MBA case studies, as well as action learning methods, such as live consulting projects and cross–national student collaborations. These active learning approaches can help students address ethical, social, and environmental issues in complex business situations.

Growth mixture models (GMMs) with nonignorable missing data have drawn increasing attention in research communities but have not been fully studied. The goal of this article is to propose and to evaluate a Bayesian method to estimate the GMMs with latent class dependent missing data. An extended GMM is first presented in which class probabilities depend on some observed explanatory variables and data missingness depends on both the explanatory variables and a latent class variable. A full Bayesian method is then proposed to estimate the model. Through the data augmentation method, conditional posterior distributions for all model parameters and missing data are obtained. A Gibbs sampling procedure is then used to generate Markov chains of model parameters for statistical inference. The application of the model and the method is first demonstrated through the analysis of mathematical ability growth data from the National Longitudinal Survey of Youth 1997 (Bureau of Labor Statistics, U.S. Department of Labor, 1997). A simulation study considering 3 main factors (the sample size, the class probability, and the missing data mechanism) is then conducted and the results show that the proposed Bayesian estimation approach performs very well under the studied conditions. Finally, some implications of this study, including the misspecified missingness mechanism, the sample size, the sensitivity of the model, the number of latent classes, the model comparison, and the future directions of the approach, are discussed.