Using upper bounds on attainable discrimination to select discrete valued features

Selection of features that will permit accurate pattern classification is a difficult task. However, if a particular data set is represented by discrete valued features, it becomes possible to determine empirically the contribution that each feature makes to the discrimination between classes. This paper extends the discrimination bound method so that both the maximum and average discrimination expected on unseen test data can be estimated. These estimation techniques are the basis of a backwards elimination algorithm that can be use to rank features in order of their discriminative power. Two problems are used to demonstrate this feature selection process: classification of the Mushroom Database, and a real-world, pregnancy related medical risk prediction task - assessment of risk of perinatal death.

Feature selection using expected attainable discrimination

We propose expected attainable discrimination (EAD) as a measure to select discrete valued features for reliable discrimination between two classes of data. EAD is an average of the area under the ROC curves obtained when a simple histogram probability density model is trained and tested on many random partitions of a data set. EAD can be incorporated into various stepwise search methods to determine promising subsets of features, particularly when misclassification costs are difficult or impossible to specify. Experimental application to the problem of risk prediction in pregnancy is described.

Psychological essentialism, implicit theories, and intergroup relations

Research on implicit person theories shows that beliefs about the malleability of human attributes have important implications for social cognition, interpersonal behavior, and intergroup relations. We argue that these implications can be understood within the framework of psychological essentialism, which extends work on implicit theories in promising directions. We review evidence that immutability beliefs covary with a broader set of essentialist beliefs, and that these essentialist beliefs are associated with stereotyping and prejudice. We then present recent studies indicating that associations between implicit person theories and stereotyping may be explained in terms of essentialist beliefs, implying a significant role for these beliefs in the psychology of group perception. Finally, we propose ways in which research and theory on essentialist beliefs might clarify and advance research on implicit person theories.

Futile treatment in hospital: Doctors’ intergroup language

Treatment that will not provide significant net benefit at the end of a person’s life (called futile treatment) is considered by many people to represent a major problem in the health sector, as it can waste resources and raise significant ethical issues. Medical treatment at the end of life involves a complex negotiation that implicates intergroup communication between health professionals, patients, and families, as well as between groups of health professionals. This study, framed by intergroup language theory, analyzed data from a larger project on futile treatment, in order to examine the intergroup language associated with futile treatment. Hospital doctors (N = 96) were interviewed about their understanding of treatment given to adult patients at the end of life that they considered futile. We conducted a discourse analysis on doctors’ descriptions of futile treatment provided by themselves and their in-group and out-group colleagues. Results pointed to an intergroup context, with patients, families, and colleagues as out-groups. In their descriptions, doctors justified their own decisions using the language of logic, ethics, and respect. Patients and families, however, were characterized in terms of wishing and wanting, as were outgroup colleagues. In addition, out-group doctors were described in strongly negative intergroup language.

Developments of the total entropy utility function for the dual purpose of model discrimination and parameter estimation in Bayesian design

The total entropy utility function is considered for the dual purpose of Bayesian design for model discrimination and parameter estimation. A sequential design setting is proposed where it is shown how to efficiently estimate the total entropy utility for a wide variety of data types. Utility estimation relies on forming particle approximations to a number of intractable integrals which is afforded by the use of the sequential Monte Carlo algorithm for Bayesian inference. A number of motivating examples are considered for demonstrating the performance of total entropy in comparison to utilities for model discrimination and parameter estimation. The results suggest that the total entropy utility selects designs which are efficient under both experimental goals with little compromise in achieving either goal. As such, the total entropy utility is advocated as a general utility for Bayesian design in the presence of model uncertainty.

Multiple alignment and sorting of peptides derived from phage-displayed random peptide libraries with polyclonal sera allows discrimination of relevant phagotopes

Biopanning of phage-displayed random peptide libraries is a powerful technique for identifying peptides that mimic epitopes (mimotopes) for monoclonal antibodies (mAbs). However, peptides derived using polyclonal antisera may represent epitopes for a diverse range of antibodies. Hence following screening of phage libraries with polyclonal antisera, including autoimmune disease sera, a procedure is required to distinguish relevant from irrelevant phagotopes. We therefore applied the multiple sequence alignment algorithm PILEUP together with a matrix for scoring amino acid substitutions based on physicochemical properties to generate guide trees depicting relatedness of selected peptides. A random heptapeptide library was biopanned nine times using no selecting antibodies, immunoglobulin G (IgG) from sera of subjects with autoimmune diseases (primary biliary cirrhosis (PBC) and type 1 diabetes) and three murine ascites fluids that contained mAbs to overlapping epitope(s) on the Ross River Virus envelope protein 2. Peptides randomly sampled from the library were distributed throughout the guide tree of the total set of peptides whilst many of the peptides derived in the absence of selecting antibody aligned to a single cluster. Moreover peptides selected by different sources of IgG aligned to separate clusters, each with a different amino acid motif. These alignments were validated by testing all of the 53 phagotopes derived using IgG from PBC sera for reactivity by capture ELISA with antibodies affinity purified on the E2 subunit of the pyruvate dehydrogenase complex (PDC-E2), the major autoantigen in PBC: only those phagotopes that aligned to PBC-associated clusters were reactive. Hence the multiple sequence alignment procedure discriminates relevant from irrelevant phagotopes and thus a major difficulty with biopanning phage-displayed random peptide libraries with polyclonal antibodies is surmounted.

Male nursing students in training: A comparative analysis of hidden advantage and discrimination

Aim The aim of this study was to examine the lived experience of men training to be registered nurses within a regional New Zealand context. Design This study draws upon the key principles of descriptive phenomenology. Sample Five male students enrolled from the 1st and 3rd year of the BN programme. Findings - A Career with Prospects - Gender inequality by superiors; - Developing professional boundaries with female colleagues; - Being unique has its advantages.