Fast kernel classifier construction using orthogonal forward selection to minimise leave-one-out misclassification rate

We propose a simple yet computationally efficient construction algorithm for two-class kernel classifiers. In order to optimise classifier's generalisation capability, an orthogonal forward selection procedure is used to select kernels one by one by minimising the leave-one-out (LOO) misclassification rate directly. It is shown that the computation of the LOO misclassification rate is very efficient owing to orthogonalisation. Examples are used to demonstrate that the proposed algorithm is a viable alternative to construct sparse two-class kernel classifiers in terms of performance and computational efficiency.

A fast linear-in-the-parameters classifier construction algorithm using orthogonal forward selection to minimize leave-one-out misclassification rate

We propose a simple and computationally efficient construction algorithm for two class linear-in-the-parameters classifiers. In order to optimize model generalization, a forward orthogonal selection (OFS) procedure is used for minimizing the leave-one-out (LOO) misclassification rate directly. An analytic formula and a set of forward recursive updating formula of the LOO misclassification rate are developed and applied in the proposed algorithm. Numerical examples are used to demonstrate that the proposed algorithm is an excellent alternative approach to construct sparse two class classifiers in terms of performance and computational efficiency.

A radial basis function network classifier to maximise leave-one-out mutual information

tWe develop an orthogonal forward selection (OFS) approach to construct radial basis function (RBF)network classifiers for two-class problems. Our approach integrates several concepts in probabilisticmodelling, including cross validation, mutual information and Bayesian hyperparameter fitting. At eachstage of the OFS procedure, one model term is selected by maximising the leave-one-out mutual infor-mation (LOOMI) between the classifier’s predicted class labels and the true class labels. We derive theformula of LOOMI within the OFS framework so that the LOOMI can be evaluated efficiently for modelterm selection. Furthermore, a Bayesian procedure of hyperparameter fitting is also integrated into theeach stage of the OFS to infer the l2-norm based local regularisation parameter from the data. Since eachforward stage is effectively fitting of a one-variable model, this task is very fast. The classifier construc-tion procedure is automatically terminated without the need of using additional stopping criterion toyield very sparse RBF classifiers with excellent classification generalisation performance, which is par-ticular useful for the noisy data sets with highly overlapping class distribution. A number of benchmarkexamples are employed to demonstrate the effectiveness of our proposed approach.

Teachers' sick leave due to mental and behavioral disorders and return to work

This manuscript presents a review of the literature about medical leaves due to mental and behavioral disorders and return to work of teachers. There are scarce published manuscripts. Most articles relate with prevalence of mental disorders and factors associated with the work organization, and did not mention intervention proposals and or changes in the work organization and teaching work. Proposed actions are discussed.

Novel screening tools for latent tuberculosis: time to leave an old friend?

PURPOSE OF REVIEW: Therapeutic inhibition of tumour necrosis factor-alpha strongly increases the risk of reactivation in latent tuberculosis infection. Recent blood tests based on antigen-specific T cell response and measuring production of interferon-gamma, so called interferon-gamma release assays (IGRAs), are promising novel tools to identify infected patients. The performance of diagnostic testing for latent tuberculosis infection in patients with rheumatic diseases will be discussed. RECENT FINDINGS: In patients with rheumatoid arthritis, IGRAs are more sensitive and more specific than traditional tuberculin skin testing. They are unaffected by Bacillus-Calmette-Guérin vaccination and most nontuberculous mycobacteria. Most comparative studies show a better performance of the IGRAs than tuberculin skin testing in terms of a higher specificity. The rate of indeterminate results may be affected by glucocorticoids and the underlying disease but appears independent of disease-modifying antirheumatic drugs. Despite using identical Mycobacterium tuberculosis antigens, the two commercially available tests show differences in clinical performance. SUMMARY: The current information about the performance of the tuberculin skin testing and the IGRAs in the detection of latent tuberculosis infection in patients with rheumatic diseases strongly suggest a clinically relevant advantage of the IGRAs. Their use will help to reduce overuse and underuse of preventive treatment in tumour necrosis factor inhibition.

Did John really promise Mary to leave?

This paper contrasts finite and non-finite complement constructions containing the matrix verb promise. Using data from the British National Corpus, I show that when no explicit mention is made of the promissee the non-finite form of complement is overwhelmingly preferred to its finite counterparts. The exact opposite is the case when the promissee is mentioned between the matrix verb and the complement clause. In addition, the promiser in the x promise y to infinitive construction is almost always pronominal. I suggest that these two facts, the dispreference for the to infinitive form of complement when the promissee is mentioned and the pronominal encoding of the promiser in such cases, are both related to the very rarity of this form of construction in English. Data is adduced showing that another rare construction, the so-called possessive -ing construction, also occurs with a disproportionate number of pronominal subjects. It is suggested that the preference for pronominal subjects in these constructions may be related to a wish to reduce the overall processing complexity of the predications in question.

Determinants to Leave Agriculture and Change Occupational Sector: Evidence from an Enlarged EU. Factor Markets Working Document No. 46, May 2013

The objective of this paper is to explore the determinants to leave agriculture and change occupational sector. We adopt a 3-step multivariate probit where we control for selection bias at two stages in the decisions to work and, at a later stage, exit agriculture. The analysis is based on the European Union Labour Force Survey data expanded with additional regional indicators. The main results suggest that younger individuals are more likely to leave farming activities, although the largest outflows of agricultural labour are mainly associated with the retirement of people. Self-employed and family workers are generally less likely to leave agriculture and those with low levels of educations are found to be significantly constrained in entering the non-farm economy. Moreover, labour market conditions at the regional level do matter for switching occupational sector. Differences in the results among the selected new member states and the EU-15 can be explained by the diverse production structures, suggesting different capacities to release and absorb labour.