Another look at linear programming for feature selection via methods of regularization

We consider statistical procedures for feature selection defined by a family of regularization problems with convex piecewise linear loss functions and penalties of l ₁ nature. Many known statistical procedures (e.g. quantile regression and support vector machines with l ₁-norm penalty) are subsumed under this category. Computationally, the regularization problems are linear programming (LP) problems indexed by a single parameter, which are known as ‘parametric cost LP’ or ‘parametric right-hand-side LP’ in the optimization theory. Exploiting the connection with the LP theory, we lay out general algorithms, namely, the simplex algorithm and its variant for generating regularized solution paths for the feature selection problems. The significance of such algorithms is that they allow a complete exploration of the model space along the paths and provide a broad view of persistent features in the data. The implications of the general path-finding algorithms are outlined for several statistical procedures, and they are illustrated with numerical examples.     

Bayesian tests for unit root and multiple breaks

A Bayesian approach is considered for identifying sources of nonstationarity for models with a unit root and breaks. Different types of multiple breaks are allowed through crash models, changing growth models, and mixed models. All possible nonstationary models are represented by combinations of zero or nonzero parameters associated with time trends, dummy for breaks, or previous levels, for which Bayesian posterior probabilities are computed. Multiple tests based on Markov chain Monte Carlo procedures are implemented. The proposed method is applied to a real data set, the Korean GDP data set, showing a strong evidence for two breaks rather than the usual unit root or one break.     

Bayesian meta-analysis of correlation coefficients through power prior

This article proposes a Bayesian approach for meta-analysis of correlation coefficients through power prior. The primary purpose of this method is to allow meta-analytic researchers to evaluate the contribution and influence of each individual study to the estimated overall effect size though power prior. We use the relationship between high-performance work systems and financial performance as an example to illustrate how to apply this method. We also introduce free online software that can be used to conduct Bayesian meta-analysis proposed in this study. Implications and future directions are also discussed in this article.     

Economic Incentive,Embeddedness, and Social Support: A Study of Korean‐Owned Nail Salon Workers’ Rotating Credit Associations1

Much of the past research on rotating credit associations (RCAs) in the U.S. Korean community has been conducted in the context of Korean entrepreneurs’ success in small businesses. By contrast, little has been known about the significance of RCAs in the lives of Korean immigrant workers. Based on a sample of Korean female workers at Korean‐owned nail salons in the New York–New Jersey area, the first aim of this study is to address whether Korean immigrant workers, like Korean immigrant merchants, take into account RCAs as a way to save money or raise capital. Second, this study also speculates about the importance of embeddedness (Granovetter) and social capital (Portes and Sensenbrenner) views for both economic behavior and a likelihood of malfeasance by RCA participants. Lastly, this study regards RCA membership as a mechanism of social support for its participating members. Overall, the analyses provide evidence that RCA membership at nail salons leads to both economic benefit and social support for some of its participants, and that embedded networks and an accompanying sense of trust have some connection to the suppression of its members’ latent malfeasance.     

Discovering the knowledge creation process of an expert group in women-friendly policy: The policy case of Seoul City

ABSTRACT

This research explores how expert knowledge is created in the process of women-friendly policy making, based on actor network theory (ANT). To address this purpose, this study uses the “Women’s Happiness in the City of Seoul” policy initiated by the local government of Seoul as one example of policy development. Research findings demonstrate that knowledge creation in expert groups followed the four stages suggested by ANT. In addition, this study found that various types of knowledge emerged from individual experts. This research elucidates the process of knowledge creation and its meanings for women-friendly policy.     

Adolescents' Expenditure On Alcohol: A Pilot Study

Minimal research has examined the extent to which underage youth access and purchase alcohol. This pilot study investigated adolescents' expenditure on alcohol, and the circumstances in which underage youth access alcohol. A street intercept survey was used to interview 187 young people aged 13 to 17 years in shopping centres in metropolitan and rural Victoria, Australia. Seventy percent of the sample self‐reported that they drink alcohol. Young people were more likely to report that they were drinkers as their income increased. More than one‐third of the sample and over half of drinkers purchased alcohol with their own money, spending on average $22 when they last purchased alcohol. Drinkers obtained their alcohol from a variety of other sources including friends over 18 years old and their parents. This study shows that underage youth have access to and pay for alcohol even if they do not purchase it personally. Further research in this area with a larger sample is clearly warranted to inform both prevention strategies and effective practice.     

Analysis of strengths and weaknesses in the National Human Resource Development systems of BRICS

National Human Resource Development (NHRD) takes a humane and long-term perspective to the human resources of each country in terms of national growth. This research evaluated the human resource development competitiveness of Brazil, Russia, India, China, South Africa (BRICS) countries from the perspective of NHRD. This study selected Global Human Resource Competitiveness Index to assess the NHRD of BRICS. Although it is generally considered that the BRICS countries have high growth potential with large population, this analysis showed that they have a common weakness in the balanced development and use of human resources from the perspective of NHRD. In this respect, this study compares and analyses the characteristics and strength/weakness of each BRICS country. This research can inform evaluation of policies on national human resource development. Additionally, it can provide basic data for identifying the status of the human resource development system.     

Computing an effective decision making group of a society using social network analysis

Recent years have witnessed how much a decision making group can be dysfunctional due to the extreme hyperpartisanship. While partisanship is crucial for the representatives to pursue the wishes of those whom they represent for, such an extremism results in a severe gridlock in the decision making progress, and makes themselves highly inefficient. It is known that such a problem can be mitigated by having negotiators in the group. This paper investigates the potential of social network analysis techniques to choose an effective leadership group of a society such that it suffers less from the extreme hyperpartisanship. We establish three essential requirements for an effective representative group, namely Influenceability, Partisanship, and Bipartisanship. Then, we formulate the problem of finding a minimum size representative group satisfying the three requirements as the minimum connected \(k\) -core dominating set problem (MC \(k\) CDSP), and show its NP-hardness. We introduce an extension of MC \(k\) CDSP, namely MC \(k\) CDSP-C, which assumes the society has a number of sub-communities and requires at least one representative from each sub-community should be in the leadership. We also propose an approximation algorithm for a subclass of MC \(k\) CDSP with \(k=2\) , and show an \(\alpha \) -approximation algorithm of MC \(k\) CDSP can be used to obtain an \(\alpha \) -approximation algorithm of MC \(k\) CDSP-SC.     

A Model for Measuring Supplier Risk: Do Operational Capability Indicators Enhance the Prediction Accuracy of Supplier Risk?

The purpose of this study was to develop a supplier risk assessment model for buyers to estimate supplier risk. It is one of the few empirical studies that considers both operational capability indicators and financial indicators; a standard logit model with five key variables (switching cost, operating profit margin, asset turnover ratio, quality capability and technological capability) was suggested as a practical tool. This model not only enhanced the accuracy of supplier risk assessment, but also served as a core element of a new supply chain management tool, ‘supplier management at risk’. More practically, the model enables purchasing firms to assess supplier risk and take proactive measures against the estimated risk.