Improved Transferability of Data‐Driven Damage Models Through Sample Selection Bias Correction

Damage models for natural hazards are used for decision making on reducing and transferring risk. The damage estimates from these models depend on many variables and their complex sometimes nonlinear relationships with the damage. In recent years, data‐driven modeling techniques have been used to capture those relationships. The available data to build such models are often limited. Therefore, in practice it is usually necessary to transfer models to a different context. In this article, we show that this implies the samples used to build the model are often not fully representative for the situation where they need to be applied on, which leads to a “sample selection bias.” In this article, we enhance data‐driven damage models by applying methods, not previously applied to damage modeling, to correct for this bias before the machine learning (ML) models are trained. We demonstrate this with case studies on flooding in Europe, and typhoon wind damage in the Philippines. Two sample selection bias correction methods from the ML literature are applied and one of these methods is also adjusted to our problem. These three methods are combined with stochastic generation of synthetic damage data. We demonstrate that for both case studies, the sample selection bias correction techniques reduce model errors, especially for the mean bias error this reduction can be larger than 30%. The novel combination with stochastic data generation seems to enhance these techniques. This shows that sample selection bias correction methods are beneficial for damage model transfer.     

Illustrating the Value of Critical Methodologies Through Third-sector Gender Studies: A Case for Pluralism

VOLUNTAS: International Journal of Voluntary and Nonprofit Organizations - To encourage methodological pluralism in the field, this paper examines an illustrative sample of articles that apply...     

Co-constructing a Conceptual Understanding of System Enactment

Clinical Social Work Journal - System enactments are co-created phenomena characterized by confounding and emotionally charged multi-person interactions that emerge through the convergence of...     

The Need to Reconcile Concepts that Characterize Systems Facing Threats

Desirable system performance in the face of threats has been characterized by various management concepts. Through semistructured interviews with editors of journals in the fields of emergency response and systems management, a literature review, and professional judgment, we identified nine related and often interchangeably used system performance concepts: adaptability, agility, reliability, resilience, resistance, robustness, safety, security, and sustainability. A better understanding of these concepts will allow system planners to pursue management strategies best suited to their unique system dynamics and specific objectives of good performance. We analyze expert responses and review the linguistic definitions and mathematical framing of these concepts to understand their applications. We find a lack of consensus on their usage between interview subjects, but by using the mathematical framing to enrich the linguistic definitions, we formulate comparative visualizations and propose distinct definitions for the nine concepts. We present a conceptual framing to relate the concepts for management purposes.     

An Evaluation of Mechanisms of Change in Multisystemic Therapy for Juvenile Justice‐Involved Youths A Decade Following Treatment

Although researchers have identified the more immediate mechanisms of change in family‐based treatments for juvenile justice‐involved youths, it is not known whether these same mechanisms continue to prevent criminal offending into adulthood. The present study evaluated whether caregiver‐directed improvements in family relations, youth prosocial peer relations, and youth academic performance during multisystemic therapy (MST) for serious and violent juvenile offenders had an impact on young adult involvement in criminal activity and sentencing 10.2 years following treatment. The results showed that improvements in family relations were associated with reduced odds of criminal outcomes a decade later for former MST participants. Furthermore, improvements in youth prosocial peer relations and academic performance were also related to lower odds of long‐term criminal activity. These results are consistent with the underlying theory of change in family‐based treatments and demonstrate that caregivers are critical to achieving and sustaining decreased antisocial behavior for youths with serious and violent criminal histories.     

Mother–Adolescent and Father–Adolescent Relationships After Divorce: Relations with Emerging Adults’ Romantic Attachment

This longitudinal study examined relations between mother–child relationship quality and father support in adolescence (ages 15–19) and emerging adults’ (ages 24–28) romantic anxiety and avoidance in a sample that experienced parental divorce in childhood (n = 72). A significant interactive effect of mother–adolescent relationship quality and father support on romantic anxiety occurred. High levels of father support protected emerging adults from the negative effects of a low quality mother–adolescent relationship. These results highlight the importance of examining effects of one parental relationship in the context of the other and the role that nonresidential fathers play in influencing their offspring’s romantic attachment.     

A reflected feature space for CART

We present an algorithm for learning oblique decision trees, called HHCART(G). Our decision tree combines learning concepts from two classification trees, HHCART and Geometric Decision Tree (GDT). HHCART(G) is a simplified HHCART algorithm that uses linear structure in the training examples, captured by a modified GDT angle bisector, to define splitting directions. At each node, we reflect the training examples with respect to the modified angle bisector to align this linear structure with the coordinate axes. Searching axis parallel splits in this reflected feature space provides an efficient and effective way of finding oblique splits in the original feature space. Our method is much simpler than HHCART because it only considers one reflected feature space for node splitting. HHCART considers multiple reflected feature spaces for node splitting making it more computationally intensive to build. Experimental results show that HHCART(G) is an effective classifier, producing compact trees with similar or better results than several other decision trees, including GDT and HHCART trees.     

The Implicit Curriculum: Student Engagement and the Role of Social Media

This article presents results from an exploratory study conducted as a component of an assessment of the implicit curriculum in one master’s-level social work program. Web-based surveys were used to collect data from 80 graduate social work students in online and face-to-face programs. Findings indicate students who perceived social media as more useful reported higher levels of engagement and participation in program governance. The implications for social work education deal with the ability to competently use social media as a tool to increase student engagement. Students will only be exposed to elements of the implicit curriculum if they are active participants in their academic environments.     

Symbolic correlation integral

This paper aims to introduce the concept of symbolic correlation integral SC that is extensively used in many scientific fields. The new correlation integral SC avoids the noisy parameter 𝜀 of the classical correlation integral, defined by Grassberger and Procaccia (1983 Grassberger, P., Procaccia, I. (1983). Measuring the strangeness of strange attractors. Physica D: Nonlinear Phenomena 9(1–2):189–208.[Crossref], [Web of Science ®] , [Google Scholar]) and extensively used for constructing correlation-integral-based statistics, as in the BDS test. Once the free parameter 𝜀 disappears, it is possible to construct a nonparametric powerful test for independence that can also be used as a diagnostic tool for model selection. The symbolic correlation integral is also extended to deal with multivariate models, and a test for causality is proposed as an example of the theoretical power of the new concept. With extensive Monte Carlo simulations, the paper shows the good size and power performance of symbolic correlation-integral-based tests.