Machine learning and design of experiments with an application to product innovation in the chemical industry

Industrial statistics plays a major role in the areas of both quality management and innovation. However, existing methodologies must be integrated with the latest tools from the field of Artificial Intelligence. To this end, a background on the joint application of Design of Experiments (DOE) and Machine Learning (ML) methodologies in industrial settings is presented here, along with a case study from the chemical industry. A DOE study is used to collect data, and two ML models are applied to predict responses which performance show an advantage over the traditional modeling approach. Emphasis is placed on causal investigation and quantification of prediction uncertainty, as these are crucial for an assessment of the goodness and robustness of the models developed. Within the scope of the case study, the models learned can be implemented in a semi-automatic system that can assist practitioners who are inexperienced in data analysis in the process of new product development.     

Postdisaster Housing Stages: A Markov Chain Approach to Model Sequences and Duration Based on Social Vulnerability

Housing recovery is an unequal and complex process presumed to occur in four stages: emergency shelter, temporary shelter, temporary housing, and permanent housing. This work questions the four-stage typology and examines how different types of shelter align with multiple housing recovery stages given different levels of social vulnerability. This article also presents a Markov chain model of the postdisaster housing recovery process that focuses on the experience of the household. The model predicts the sequence and timing of a household going through housing recovery, capturing households that end in either permanent housing or a fifth possible stage of failure. The probability of a household transitioning through the stages is computed using a transition probability matrix (TPM). The TPM is assembled using proposed transition probability models that vary with the social vulnerability of the household. Monte Carlo techniques are applied to demonstrate the range of sequences and timing that households experience going through the housing recovery process. A set of computational rules are established for sending a household to the fifth stage, representing a household languishing in unstable housing. This predictive model is exemplified on a virtual community, Centerville, where following a severe earthquake scenario, differences in housing recovery times exceed four years. The Centerville analysis results in nearly 5% of households languishing in unstable housing, thereby failing to reach housing recovery. These findings highlight the disparate trajectories experienced by households with different levels of social vulnerability. Recommendations are provided at the end for more equitable postdisaster recovery policies.     

Managing Logistics in Regional Banknote Supply Chain Under Security Concerns

We study the logistics problem faced by Regional Branches (RBs) of a central bank in managing the currency supply under security concerns. While making banknote supply decisions to Sub‐Branches (SBs), the management of RB must achieve two goals simultaneously: (i) guarantee that each SB has sufficient inventories of all denominations of banknotes to satisfy the demands from all commercial banks within its service area, and (ii) control the annual spending on this banknote supply operation. Due to security concerns, the following methods are implemented in the process of transporting banknotes: (i) the capacity of a cash truck is limited by the total face value (instead of the physical space) of banknotes, and (ii) empty decoy trucks are deployed along with the trucks filled with banknotes. After deriving a polynomial‐time strategy to guarantee an optimal solution for the special Bin‐Packing Problem faced in this study, we provide an exact formulation for the RB's supply planning problem. We also propose several polynomial‐time algorithms for deriving either optimal or near‐optimal solutions for the problem under different settings. Using the weekly demand data obtained from the central bank, we verify the performance of our algorithms, and analyze the impacts of changes in these features and in the fleet capacity on the total cost incurred by an RB under various scenarios.     

Conceptualizing Childbearing Ambivalence: A Social and Dynamic Perspective

Childbearing ambivalence is often conceptualized as a state of conflicting desires about having a child that is characteristic of particular individuals or life stages. This study proposes that childbearing ambivalence is dynamic and situational, resulting from the multiple sociocultural frames surrounding childbearing. Using eight waves of prospective data from a population‐based sample of young adults in Malawi, results show that 41% of women and 48% of men are ambivalent about childbearing at some point in the 2.5‐year study. There is limited evidence that ambivalence is related to individual sociodemographic or psychosocial characteristics aside from gender; rather, ambivalence is tied to life course markers such as school enrollment and age. In addition, life course transitions and changes in relationships, health, and economic factors are associated with the onset of ambivalence, supporting the theory that ambivalence is a dynamic state that men and women frequently pass through as their lives and circumstances change.     

Framing the Women’s March on Washington: Media coverage and organizational messaging alignment

Half a million citizens participated in the Women’s March on Washington the day after President Trump’s inauguration, starting a political movement. The march communicated key messages to the public directly and via the media. This study explores how media coverage framed those key messages through content analysis. Media frames mentioned all key messages, emphasizing solidarity and activation at the grassroots, and in a way that both supported and challenged organizational messages. Theoretical and practical implications are discussed.     

Approximate Bayesian inference for latent Gaussian models by using integrated nested Laplace approximations

Summary.  Structured additive regression models are perhaps the most commonly used class of models in statistical applications. It includes, among others, (generalized) linear models, (generalized) additive models, smoothing spline models, state space models, semiparametric regression, spatial and spatiotemporal models, log-Gaussian Cox processes and geostatistical and geoadditive models. We consider approximate Bayesian inference in a popular subset of structured additive regression models, latent Gaussian models , where the latent field is Gaussian, controlled by a few hyperparameters and with non-Gaussian response variables. The posterior marginals are not available in closed form owing to the non-Gaussian response variables. For such models, Markov chain Monte Carlo methods can be implemented, but they are not without problems, in terms of both convergence and computational time. In some practical applications, the extent of these problems is such that Markov chain Monte Carlo sampling is simply not an appropriate tool for routine analysis. We show that, by using an integrated nested Laplace approximation and its simplified version, we can directly compute very accurate approximations to the posterior marginals. The main benefit of these approximations is computational: where Markov chain Monte Carlo algorithms need hours or days to run, our approximations provide more precise estimates in seconds or minutes. Another advantage with our approach is its generality, which makes it possible to perform Bayesian analysis in an automatic, streamlined way, and to compute model comparison criteria and various predictive measures so that models can be compared and the model under study can be challenged.     

Stakeholder perceptions of contextual factors related to PBIS implementation in high need schools

Positive behavior intervention and supports (PBIS) is a thoroughly researched three-tiered framework for preventing and treating challenging behavior in schools. While more than 20 years of research contribute to the literature base, less is known about PBIS implementation in high-needs school settings, particularly those that serve students predominantly from low-income families and majority minority student populations across urban and rural districts. This study investigated the critical contextual factors of PBIS implementation from the stakeholder perspective. Specifically, we surveyed 161 participants including teachers, school staff members, and administrators who were PBIS team members and had received technical assistance in PBIS implementation. The completed surveys were analyzed with participants and their schools coded for district type, Title I funding, and position in order to determine how the perceptions of issues such as buy-in, administrator support, PBIS in secondary school settings, and poverty affect PBIS implementation and team needs. The factor analysis and resulting structure along with statistically significant results for each factor are presented along with implications.     

Recent Social Work Practitioners’ Understanding and Use of Evidence-Based Practice and Empirically Supported Treatments

The aim of this study was to evaluate how CSWE’s 2008 shift placing more emphasis on research have affected newly trained social workers’ use of evidence-based practice (EBP). This qualitative study examined the educational and practice experiences of newly trained social workers and how those experiences influence the use of EBP and empirically supported interventions/treatments in the field. Thirteen newly graduated social workers were interviewed using a semi-structured interview. Twenty-three codes emerged from the interviews, with the most prominent theme being a sense of overall confusion about EBP. Other themes included lack of educational preparation, lack of agency resources, and prohibitive agency culture. Implications for social work education and practice are discussed.     

The Mental Health Advantage of Immigrant‐Background Youth: The Role of Family Factors

Children of immigrant background, despite problems with acculturation, poverty, and discrimination, have better mental health than children of native parents. We asked whether this is a result of immigrant families' characteristics such as family structure and relations. Using a new comparative study on the integration of immigrant‐background youth conducted in England, Germany, the Netherlands, and Sweden (N = 18,716), particularly strong associations with mental health (internalizing and externalizing problems) were found for family structure, family cohesion, and parental warmth. Overall, half of the advantage in internalizing and externalizing problems among immigrant‐background youth could be accounted for by our measures of family structure and family relations, with family cohesion being particularly important.