Number Of Latent Dimensions Research Articles

Translation, cultural adaptation and pilot testing of a questionnaire measuring the factors affecting the acceptance of telemedicine by Greek cancer patients.

The emergency presented through the COVID-19 pandemic exposed the need to adopt remote, technology-driven solutions and make healthcare services more resilient. To do so, we need technological applications (i.e., telemedicine) that are designed and tailored to the end-users (i.e., chronic patients) needs and the type of healthcare service they get (i.e., cancer care). The requirements above are especially relevant to Greece, being a country with numerous sparsely populated regions (e.g., islands, regions at the borders) and a deteriorating access to healthcare for all citizens. Trying to address such diverse problems and needs, there have been multiple, different telemedicine and telecare projects in Greece in the past years. To support the future design and implementation of such endeavours, in this study we translated a questionnaire measuring the acceptance of telemedicine by patients and adapted it to the Greek context. We continued by running a small-scale pilot with 73 Greek women with breast cancer to assess the adapted instrument for its reliability and construct validity. The created questionnaire had good overall and internal reliability scores for most sub-scales. Factor analysis did not identify the same number of latent dimensions as the original theoretical model. Reverse wording items needing to be recoded were identified, and items that could be omitted in future versions of the questionnaire. Increasing the sample size for the purposes of a longitudinal study, the construct, convergent, and discriminant validity are elements to be further examined in future studies. It is envisaged that the creation of this questionnaire will support the adoption of telemedicine by Greek healthcare services into more routine areas of patient care provision.

The Role of Dimensionality Reduction in Classification

Dimensionality reduction (DR) is often used as a preprocessing step in classification, but usually one first fixes the DR mapping, possibly using label information, and then learns a classifier (a filter approach). Best performance would be obtained by optimizing the classification error jointly over DR mapping and classifier (a wrapper approach), but this is a difficult nonconvex problem, particularly with nonlinear DR. Using the method of auxiliary coordinates, we give a simple, efficient algorithm to train a combination of nonlinear DR and a classifier, and apply it to a RBF mapping with a linear SVM. This alternates steps where we train the RBF mapping and a linear SVM as usual regression and classification, respectively, with a closed-form step that coordinates both. The resulting nonlinear low-dimensional classifier achieves classification errors competitive with the state-of-the-art but is fast at training and testing, and allows the user to trade off runtime for classification accuracy easily. We then study the role of nonlinear DR in linear classification, and the interplay between the DR mapping, the number of latent dimensions and the number of classes. When trained jointly, the DR mapping takes an extreme role in eliminating variation: it tends to collapse classes in latent space, erasing all manifold structure, and lay out class centroids so they are linearly separable with maximum margin.

Bilingual idiosyncratic dimensions of language attitudes

The goal of this study is to identify new dimensions of language attitudes to allow for both their multidimensionality and possible language-specificity stemming from local sociolinguistic environments. Adopting a two-step methodology comprising (1) elicitation of adjectives in group interviews and (2) employment of the semantic differential technique within a direct approach, this article demonstrates that language attitudes of bilinguals may be made up of a number of latent dimensions that go beyond those found in previous academic studies. In particular, Italian English bilinguals in Australia rate their languages according to three idiosyncratic dimensions only partly ascertained in the literature: attractiveness, superiority and efficiency. These three dimensions, emerged through rotated principal component analysis, reveal the significance of bilingualism in attitude formation. Moreover, this study provides insights on language attitudes as constructions avulsed from their contextualised manifestations and indeed accounts for both their language-specific singularity and intrinsic multidimensionality.

High-dimensional maximum marginal likelihood item factor analysis by adaptive quadrature

Although the Bock–Aitkin likelihood-based estimation method for factor analysis of dichotomous item response data has important advantages over classical analysis of item tetrachoric correlations, a serious limitation of the method is its reliance on fixed-point Gauss-Hermite (G-H) quadrature in the solution of the likelihood equations and likelihood-ratio tests. When the number of latent dimensions is large, computational considerations require that the number of quadrature points per dimension be few. But with large numbers of items, the dispersion of the likelihood, given the response pattern, becomes so small that the likelihood cannot be accurately evaluated with the sparse fixed points in the latent space. In this paper, we demonstrate that substantial improvement in accuracy can be obtained by adapting the quadrature points to the location and dispersion of the likelihood surfaces corresponding to each distinct pattern in the data. In particular, we show that adaptive G-H quadrature, combined with mean and covariance adjustments at each iteration of an EM algorithm, produces an accurate fast-converging solution with as few as two points per dimension. Evaluations of this method with simulated data are shown to yield accurate recovery of the generating factor loadings for models of upto eight dimensions. Unlike an earlier application of adaptive Gibbs sampling to this problem by Meng and Schilling, the simulations also confirm the validity of the present method in calculating likelihood-ratio chi-square statistics for determining the number of factors required in the model. Finally, we apply the method to a sample of real data from a test of teacher qualifications.

STOCHASTIC MODELS AND FACTOR ANALYSIS

The present paper discusses the relationship between the traditional factor analytic model and stochastic models of test behaviour for the case of dichotomous manifest variables. It is shown that the factor loadings can be interpreted as parameters of item trace lines. Furthermore, by comparing a linear and a nonlinear model of item characteristics, it is suggested that the assumption of a small number of latent dimensions or factors (small in comparison to the number of manifest variables) always requires the estimation of communalities making the explicit conceptualization of specific factors unnecessary. In addition, it is indicated that, because of the arbitrariness of criteria for the number of common factors, the general tendency among researchers to extract a small number of latent dimensions will, for a given data set, lead to quite similar results regardless of the trace line model chosen.