Testimator of the scale parameter of the exponential distribution using linex loss function

The present paper investigates the properties of a testimator of scale of an exponential distribution under Linex loss function. The risk function of testimator is derived and compared with that of an admissible estimator relative to Linex loss function. The shrinkage testimator is proposed which is the extension of testimator and its properties have been discussed. The level of significance of testimator is decided on the basis of Akaike information criterion following Hirano (1977, 1978). It is found that the testimator and shrinkage testimator dominates the admissible estimator in terms of risk in certain parametric space.     

A noise-based stabilizer for convolutional neural networks

Overfitting occurs when one tries to train a large model on small amount of data. Regularizing a neural network using prior knowledge remains a topic of research as it is not concluded how much prior information can be given to the neural network. In this paper, a novel algorithm is introduced which uses regularization to train a neural network without increasing the dataset. A trivial prior information of a class label is supplied to the model while training. Laplace noise is introduced to the intermediate layer for more generalization. The results show significant improvement in accuracy on the standard datasets for a simple Convolutional Neural Network (CNN). While the proposed method outperforms previous regularization techniques like dropout and batch normalization, it can also be applied with them for further improvement in the performance. On the variants of MNIST, proposed algorithm achieved an average 48% increment in the test accuracy.     

Optimal design and maintenance of a repairable multi-state system with standby components

The configuration of a repairable system directly influences its performance measures, such as mean time between failures and steady state availability. Additionally, maintenance strategies such as corrective, preventive, and condition-based can affect the performance of the system. The objective of this work is to investigate the trade-offs between the configuration of a repairable multi-state system with binary components and its maintenance strategy. The corresponding stochastic process for the proposed model is formulated using the continuous-time Markov process and important performance measures of such a multi-state system are derived. An optimization model is introduced for the cost-effective design of this repairable multi-state system. The results are demonstrated using a numerical example for a power generation system.     

Food subsidy,income transfer and the poor: A comparative analysis of the public distribution system in India's states

The Targeted Public Distribution System (TPDS) – the largest food subsidy program in India – has been a dismal failure in targeting the poor. The present paper examines its performance in three Indian states – Andhra Pradesh, Maharashtra and Rajasthan, based on primary data collected for this study. As real income transfers through food price subsidies are a tiny fraction of expenditure per person under this scheme, a contribution of the present study is to model determinants of real income transfers through subsidised wheat, rice and sugar. The analysis throws new light on how income transfers vary with economic status of a household, inequality in the distribution of land in a village, amount of food price subsidy, transaction costs of buying from ‘fair price shops’ (FPS), and supply shortages. The policy implications of these results are profound. Desperate measures such as a universal food subsidy enshrined in a proposed National Food Security Act are rejected on the grounds that the enormous leakages and wastage under the present TPDS would only get worse. Assertions that a universal food subsidy is the only option consistent with the right to food are rejected as mistaken. Instead, it is worthwhile to overhaul the PDS within the existing outlay.     

Approximate prediction limit for weibull failure based on preliminary test estimator

In this paper we have considered type II censored sample from a two parameter Weibull distribution with the known scale parameter. Using the preliminary test estimator of the unknown shape parameter (3 proposed by Pandey (1983), the paper derives a method of finding the approximate prediction limit for the minimum or, more generally,the j^th smallest of a set of future observations from the Weibull or even extreme-value distribution     

Use of social impact bonds to address social problems: Understanding contractual risks and transaction costs

Social impact bonds, a recent innovation in social finance, are designed to harness capital and knowledge from private nonprofit, for‐profit, and public entities to address pressing social problems. Although there is great policy interest in understanding how social impact bonds can be used to tackle social problems, the emergent nature of social impact bonds makes it hard to find relevant data and evidence. To overcome this challenge, we use single‐significant‐case sampling as our research design strategy. We conduct an in‐depth case study of the Social Innovation Financing Youth Recidivism Project in Massachusetts. Our case study is comprised of a qualitative analysis of the multiparty contract and multiyear quantitative benefit–cost analysis to understand transaction costs. We draw upon contract theory to develop an analytical framework for the case analysis and highlight the risks and safeguards for the various parties to the contract, and conduct a formal benefit–cost analysis to map out transaction costs. We conclude with a discussion of study implications and future research.     

Connecting older grandmothers raising grandchildren with community resources improves family resiliency,social support,and caregiver self-efficacy

Custodial grandparenting can be especially challenging for older grandmothers facing age specific issues. Kinship navigator programs are social service delivery programs intended to inform grandparents and other relatives raising children about available resources and services, provide information specific to their individual needs, and help families navigate service systems. Our study utilizes self-report data from one kinship navigator federal demonstration project, which used a randomized control trial, to examine demographic characteristics for grandmothers under and over 55 years of age, whether grandmother caregivers (≥55 years) improve family resilience, social support, and caregiver self-efficacy, and which interventions improved outcomes for grandmothers (≥55 years). Each participant was randomly assigned to one of four groups: Usual Care (traditional child welfare services), Standard Care (family support and case management), Peer-to-Peer Care Only, and Full Kin Tech Care (peer navigators with computer access and interdisciplinary team). Thirty-nine percent of grandmothers (55-75 years) were mostly living in poverty, predominantly Caucasian, with 36% identifying as African American/Black, with at least one to two children at home. Repeated-measures ANOVAs for each subscale showed statistically significant within- and between-group differences for Family Functioning, Social Supports, Concrete Supports, Child Development, and Nurturing and Attachment, with the exception of Usual Care, which showed a decline in protective factors consistently across subscales. Future research with kinship families could qualitatively examine the experiences for older women in navigator programs and replication of kinship navigator programs could build capacity in data collection and maintenance systems to gain better perspective about how systems of care impact families.     

THE AGRICULTURAL PRODUCTIVITY GAP IN EUROPE

For 15 European countries over the 1970–2004 period we find large and persistent agricultural productivity gaps, the ratio of value added per hour in nonagriculture to that in agriculture. Comparing the gap in value added per hour to the wage gap between the two sectors suggests that value added in the data is mismeasured. We further find that, controlling for differences in gross domestic product per capita and institutions, the mismeasurement is positively related to self‐employed share of hours in agriculture. Correcting for underreporting of self‐employment income using our preferred correction factor reduces the measured agricultural productivity gap by 38%. These findings suggest that underreporting can account for a significant portion of the measured agricultural productivity gap. (JEL E01, O47, O52, Q10)