			Home About us Contact

Latent Variable Models (latent + variable_models)

Distribution by Scientific Domains

Mathematics and Statistics	50%

Selected Abstracts

Understanding and using the implicit association test: V. measuring semantic aspects of trait self-concepts

EUROPEAN JOURNAL OF PERSONALITY, Issue 8 2008
Konrad Schnabel
Abstract Implicit Association Tests (IATs) often reveal strong associations of self with positive rather than negative attributes. This poses a problem in using the IAT to measure associations involving traits with either positive or negative evaluative content. In two studies, we employed non-bipolar but evaluatively balanced Big Five traits as attribute contrasts and explored correlations of IATs with positive (e.g. sociable vs. conscientious) or negative (e.g. reserved vs. chaotic) attributes. Results showed (a) satisfactory internal consistencies for all IATs, (b) explicit,explicit and implicit,implicit correlations that were moderate to high and comparable in strength after both were corrected for attenuation and (c) better model fit for latent variable models that linked the implicit and explicit measures to distinct latent factors rather to the same factor. Together, the results suggest that IATs can validly assess the semantic aspect of trait self-concepts and that implicit and explicit self-representations are, although correlated, also distinct constructs. Copyright © 2008 John Wiley & Sons, Ltd. [source]

On-line expectation,maximization algorithm for latent data models

JOURNAL OF THE ROYAL STATISTICAL SOCIETY: SERIES B (STATISTICAL METHODOLOGY), Issue 3 2009
Olivier Cappé
Summary., We propose a generic on-line (also sometimes called adaptive or recursive) version of the expectation,maximization (EM) algorithm applicable to latent variable models of independent observations. Compared with the algorithm of Titterington, this approach is more directly connected to the usual EM algorithm and does not rely on integration with respect to the complete-data distribution. The resulting algorithm is usually simpler and is shown to achieve convergence to the stationary points of the Kullback,Leibler divergence between the marginal distribution of the observation and the model distribution at the optimal rate, i.e. that of the maximum likelihood estimator. In addition, the approach proposed is also suitable for conditional (or regression) models, as illustrated in the case of the mixture of linear regressions model. [source]

Estimating Returns on Commercial Real Estate: A New Methodology Using Latent-Variable Models

REAL ESTATE ECONOMICS, Issue 2 2000
David C. Ling
Despite their widespreao use as benchmarks of U.S. commercial real estate returns, indexes produced by the National Council of Real Estate Investment Fiduciaries (NCREIF) are subject to measurement problems that severely impair their ability to capture the true risk,return characteristics,especially volatility,of privately held commercial real estate. We utilize latent-variable statistical methods to estimate an alternative index of privately held (unsecuritized) commercial real estate returns. Latent-variable methods have been extensively applied in the behavioral sciences and, more recently, in finance and economics. Unlike factor analysis or other unconditional statistical approaches, latent variable models allow us to extract interpretable common information about unobserved private real estate returns using the information contained in various competing measures of returns that are measured with error. We find that our latent-variable real estate return series is approximately twice as volatile as the aggregate NCREIF total return index, but less than half as volatile as the NAREIT equity index. Overall, our results strongly support the use of latent-variable statistical models in the construction of return series for commercial real estate. [source]

Marginalized Models for Moderate to Long Series of Longitudinal Binary Response Data

BIOMETRICS, Issue 2 2007
Jonathan S. Schildcrout
Summary Marginalized models (Heagerty, 1999, Biometrics55, 688,698) permit likelihood-based inference when interest lies in marginal regression models for longitudinal binary response data. Two such models are the marginalized transition and marginalized latent variable models. The former captures within-subject serial dependence among repeated measurements with transition model terms while the latter assumes exchangeable or nondiminishing response dependence using random intercepts. In this article, we extend the class of marginalized models by proposing a single unifying model that describes both serial and long-range dependence. This model will be particularly useful in longitudinal analyses with a moderate to large number of repeated measurements per subject, where both serial and exchangeable forms of response correlation can be identified. We describe maximum likelihood and Bayesian approaches toward parameter estimation and inference, and we study the large sample operating characteristics under two types of dependence model misspecification. Data from the Madras Longitudinal Schizophrenia Study (Thara et al., 1994, Acta Psychiatrica Scandinavica90, 329,336) are analyzed. [source]