An analysis of two probabilistic model induction techniques

Two probabilistic model induction techniques, cart and constructor, are compared, via a series of experiments, in terms of their ability to induce models that are both interpretable and predictive. The experiments show that, although both algorithms are able to deliver classifiers with predictive performance close to that of the optimal Bayes rule,constructor is able to generate a probabilistic model that is more easily interpretable than the cart model. On the other hand, cart is a more mature algorithm and is capable of handling many more situations (e.g., real-valued training sets) thanconstructor. A variety of characteristics of both algorithms are compared, and suggestions for future research are made.     

Participation in Practice

This paper is one of a series written by members of the Mental Handicap Research Project at the University College of North Wales in connection with the All-Wales Strategy. Previous papers describe local arrangements for the planning of new services and discuss in general terms issues concerning partnership and participation. This paper describes modes of parental participation in the planning of new services for the mentally handicapped.     

Online surveys for BGLT research: issues and techniques

Online surveys are becoming increasingly popular for accessing less visible and decentralized populations, including bisexual, gay, lesbian and transgender (BGLT) populations. Advances in technology and convenience for the both the researcher and the participant have facilitated this trend. In this paper, we explore issues related to conducting BGLT survey research online, such as making decisions about survey formats, target populations and recruitment, compensation, access.     

Dealing with unknown variables in policy/program evaluation

A leading objection to the feasibility of applying benefit-cost analysis in evaluation is that one often cannot know the benefits, the costs, or the probabilities of their occurring for a given proposal. Threshold analysis involves handling unknown variables by converting relevant evaluation problems into questions as to whether a given benefit, cost, probability, or combination of these elements is more or less than a threshold. Above the threshold, the proposed program would be relatively profitable; and below the threshold, it would be relatively unprofitable. Applying that analysis may require the drawing of simple graphs to stimulate the minds of knowledgeable persons as to the range of the actual benefits, costs or probabilities. The approach can be applied when the benefits or the costs are either monetary or nonmonetary, and when the problems involve either go/no-go decisions or conflicting-choice decisions.     

Danish home care policy and the family: implications for the United States

This paper provides an overview of reforms in Danish long-term care initiated in the early 1980s, describes the relationship between elder care in Denmark and the family, and considers implications for U.S. policy. The success of Denmark's community-based experimentation with new models of home care and housing for the elderly resulted in a national decision to eliminate new construction of nursing homes and increase access to publicly funded home care. Lingering concern that the provision of paid assistance for the elderly could undermine family structure is allayed by the findings of a recent survey: Three-fourths of the elderly report seeing their children on a weekly or more frequent basis. Findings from the Danish experience provide evidence that community-based services can aid family caregivers, enable the frail elderly to live in the setting of their choice, and be cost-effective from a public policy perspective.     

Fisher's Hypergeometric Test for a Comparison in a Finite Population

Graphical methods have played a central role in the development of statistical theory and practice. This presentation briefly reviews some of the highlights in the historical development of statistical graphics and gives a simple taxonomy that can be used to characterize the current use of graphical methods. This taxonomy is used to describe the evolution of the use of graphics in some major statistical and related scientific journals.

Some recent advances in the use of graphical methods for statistical analysis are reviewed, and several graphical methods for the statistical presentation of data are illustrated, including the use of multicolor maps.