Backward elimination model construction for regression and classification using leave-one-out criteria

A fundamental principle in practical nonlinear data modeling is the parsimonious principle of constructing the minimal model that explains the training data well. Leave-one-out (LOO) cross validation is often used to estimate generalization errors by choosing amongst different network architectures (M. Stone, "Cross validatory choice and assessment of statistical predictions", J. R. Stast. Soc., Ser. B, 36, pp. 117-147, 1974). Based upon the minimization of LOO criteria of either the mean squares of LOO errors or the LOO misclassification rate respectively, we present two backward elimination algorithms as model post-processing procedures for regression and classification problems. The proposed backward elimination procedures exploit an orthogonalization procedure to enable the orthogonality between the subspace as spanned by the pruned model and the deleted regressor. Subsequently, it is shown that the LOO criteria used in both algorithms can be calculated via some analytic recursive formula, as derived in this contribution, without actually splitting the estimation data set so as to reduce computational expense. Compared to most other model construction methods, the proposed algorithms are advantageous in several aspects; (i) There are no tuning parameters to be optimized through an extra validation data set; (ii) The procedure is fully automatic without an additional stopping criteria; and (iii) The model structure selection is directly based on model generalization performance. The illustrative examples on regression and classification are used to demonstrate that the proposed algorithms are viable post-processing methods to prune a model to gain extra sparsity and improved generalization.

Parameter estimation based on stacked regression and evolutionary algorithms

A new parameter-estimation algorithm, which minimises the cross-validated prediction error for linear-in-the-parameter models, is proposed, based on stacked regression and an evolutionary algorithm. It is initially shown that cross-validation is very important for prediction in linear-in-the-parameter models using a criterion called the mean dispersion error (MDE). Stacked regression, which can be regarded as a sophisticated type of cross-validation, is then introduced based on an evolutionary algorithm, to produce a new parameter-estimation algorithm, which preserves the parsimony of a concise model structure that is determined using the forward orthogonal least-squares (OLS) algorithm. The PRESS prediction errors are used for cross-validation, and the sunspot and Canadian lynx time series are used to demonstrate the new algorithms.

Forecasting with difference and trend stationary models

Although difference-stationary (DS) and trend-stationary (TS) processes have been subject to considerable analysis, there are no direct comparisons for each being the data-generation process (DGP). We examine incorrect choice between these models for forecasting for both known and estimated parameters. Three sets of Monte Carlo simulations illustrate the analysis, to evaluate the biases in conventional standard errors when each model is mis-specified, compute the relative mean-square forecast errors of the two models for both DGPs, and investigate autocorrelated errors, so both models can better approximate the converse DGP. The outcomes are surprisingly different from established results.

Information customization and food choice

In this paper we employ a hypothetical discrete choice experiment (DCE) to examine how much consumers are willing to pay to use technology to customize their food shopping. We conjecture that customized information provision can aid in the composition of a healthier shop. Our results reveal that consumers are prepared to pay relatively more for individual specic information as opposed to generic nutritional information that is typically provided on food labels. In arriving at these results we have examined various model specications including those that make use of ex-post de-brieng questions on attribute nonattendance and attribute ranking information and those that consider the time taken to complete the survey. Our main results are robust to the various model specications we examine

Self-congruence, functional congruence, and destination choice

This study's purpose is to investigate the effects of self-congruence and functional congruence on tourists' destination choice. The present research contributes to the gap in the consumer behavior literature by examining the relationships among self-congruence, functional congruence, and destination choice. Based on a sample of 367 British residents, the three research hypotheses are tested using multinomial logistic regression analysis. The study results suggest that a tourist's destination choice is influenced strongly by functional congruence, but not by self-congruence. The article closes with theoretical and managerial implications as well as future research directions.

The Capm and Fama-French models in Brazil: a comparative study

This paper confronts the Capital Asset Pricing Model - CAPM - and the 3-Factor Fama-French - FF - model using both Brazilian and US stock market data for the same Sample period (1999-2007). The US data will serve only as a benchmark for comparative purposes. We use two competing econometric methods, the Generalized Method of Moments (GMM) by (Hansen, 1982) and the Iterative Nonlinear Seemingly Unrelated Regression Estimation (ITNLSUR) by Burmeister and McElroy (1988). Both methods nest other options based on the procedure by Fama-MacBeth (1973). The estimations show that the FF model fits the Brazilian data better than CAPM, however it is imprecise compared with the US analog. We argue that this is a consequence of an absence of clear-cut anomalies in Brazilian data, specially those related to firm size. The tests on the efficiency of the models - nullity of intercepts and fitting of the cross-sectional regressions - presented mixed conclusions. The tests on intercept failed to rejected the CAPM when Brazilian value-premium-wise portfolios were used, contrasting with US data, a very well documented conclusion. The ITNLSUR has estimated an economically reasonable and statistically significant market risk premium for Brazil around 6.5% per year without resorting to any particular data set aggregation. However, we could not find the same for the US data during identical period or even using a larger data set. Este estudo procura contribuir com a literatura empírica brasileira de modelos de apreçamento de ativos. Dois dos principais modelos de apreçamento são Infrontados, os modelos Capital Asset Pricing Model (CAPM)e de 3 fatores de Fama-French. São aplicadas ferramentas econométricas pouco exploradas na literatura nacional na estimação de equações de apreçamento: os métodos de GMM e ITNLSUR. Comparam-se as estimativas com as obtidas de dados americanos para o mesmo período e conclui-se que no Brasil o sucesso do modelo de Fama e French é limitado. Como subproduto da análise, (i) testa-se a presença das chamadas anomalias nos retornos, e (ii) calcula-se o prêmio de risco implícito nos retornos das ações. Os dados revelam a presença de um prêmio de valor, porém não de um prêmio de tamanho. Utilizando o método de ITNLSUR, o prêmio de risco de mercado é positivo e significativo, ao redor de 6,5% ao ano.

Short communication: Principal components and factor analytic models for test-day milk yield in Brazilian Holstein cattle

Conselho Nacional de Desenvolvimento Científico e Tecnológico (CNPq)

Comparison of a genetic group and unknown paternity models for growth traits in Nellore cattle

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Quantitative and qualitative dynamics of exotic and native blowflies ( Diptera: Calliphoridae) with migrations among municipalities

Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP)

Short communication: Principal components and factor analytic models for test-day milk yield in Brazilian Holstein cattle

A total of 46,089 individual monthly test-day (TD) milk yields (10 test-days), from 7,331 complete first lactations of Holstein cattle were analyzed. A standard multivariate analysis (MV), reduced rank analyses fitting the first 2, 3, and 4 genetic principal components (PC2, PC3, PC4), and analyses that fitted a factor analytic structure considering 2, 3, and 4 factors (FAS2, FAS3, FAS4), were carried out. The models included the random animal genetic effect and fixed effects of the contemporary groups (herd-year-month of test-day), age of cow (linear and quadratic effects), and days in milk (linear effect). The residual covariance matrix was assumed to have full rank. Moreover, 2 random regression models were applied. Variance components were estimated by restricted maximum likelihood method. The heritability estimates ranged from 0.11 to 0.24. The genetic correlation estimates between TD obtained with the PC2 model were higher than those obtained with the MV model, especially on adjacent test-days at the end of lactation close to unity. The results indicate that for the data considered in this study, only 2 principal components are required to summarize the bulk of genetic variation among the 10 traits.

Performance evaluation of low environmental impact asphalt concretes using the mechanistic empirical design method based on laboratory fatigue and permanent deformation models

The "sustainability" concept relates to the prolonging of human economic systems with as little detrimental impact on ecological systems as possible. Construction that exhibits good environmental stewardship and practices that conserve resources in a manner that allow growth and development to be sustained for the long-term without degrading the environment are indispensable in a developed society. Past, current and future advancements in asphalt as an environmentally sustainable paving material are especially important because the quantities of asphalt used annually in Europe as well as in the U.S. are large. The asphalt industry is still developing technological improvements that will reduce the environmental impact without affecting the final mechanical performance. Warm mix asphalt (WMA) is a type of asphalt mix requiring lower production temperatures compared to hot mix asphalt (HMA), while aiming to maintain the desired post construction properties of traditional HMA. Lowering the production temperature reduce the fuel usage and the production of emissions therefore and that improve conditions for workers and supports the sustainable development. Even the crumb-rubber modifier (CRM), with shredded automobile tires and used in the United States since the mid 1980s, has proven to be an environmentally friendly alternative to conventional asphalt pavement. Furthermore, the use of waste tires is not only relevant in an environmental aspect but also for the engineering properties of asphalt [Pennisi E., 1992]. This research project is aimed to demonstrate the dual value of these Asphalt Mixes in regards to the environmental and mechanical performance and to suggest a low environmental impact design procedure. In fact, the use of eco-friendly materials is the first phase towards an eco-compatible design but it cannot be the only step. The eco-compatible approach should be extended also to the design method and material characterization because only with these phases is it possible to exploit the maximum potential properties of the used materials. Appropriate asphalt concrete characterization is essential and vital for realistic performance prediction of asphalt concrete pavements. Volumetric (Mix design) and mechanical (Permanent deformation and Fatigue performance) properties are important factors to consider. Moreover, an advanced and efficient design method is necessary in order to correctly use the material. A design method such as a Mechanistic-Empirical approach, consisting of a structural model capable of predicting the state of stresses and strains within the pavement structure under the different traffic and environmental conditions, was the application of choice. In particular this study focus on the CalME and its Incremental-Recursive (I-R) procedure, based on damage models for fatigue and permanent shear strain related to the surface cracking and to the rutting respectively. It works in increments of time and, using the output from one increment, recursively, as input to the next increment, predicts the pavement conditions in terms of layer moduli, fatigue cracking, rutting and roughness. This software procedure was adopted in order to verify the mechanical properties of the study mixes and the reciprocal relationship between surface layer and pavement structure in terms of fatigue and permanent deformation with defined traffic and environmental conditions. The asphalt mixes studied were used in a pavement structure as surface layer of 60 mm thickness. The performance of the pavement was compared to the performance of the same pavement structure where different kinds of asphalt concrete were used as surface layer. In comparison to a conventional asphalt concrete, three eco-friendly materials, two warm mix asphalt and a rubberized asphalt concrete, were analyzed. The First Two Chapters summarize the necessary steps aimed to satisfy the sustainable pavement design procedure. In Chapter I the problem of asphalt pavement eco-compatible design was introduced. The low environmental impact materials such as the Warm Mix Asphalt and the Rubberized Asphalt Concrete were described in detail. In addition the value of a rational asphalt pavement design method was discussed. Chapter II underlines the importance of a deep laboratory characterization based on appropriate materials selection and performance evaluation. In Chapter III, CalME is introduced trough a specific explanation of the different equipped design approaches and specifically explaining the I-R procedure. In Chapter IV, the experimental program is presented with a explanation of test laboratory devices adopted. The Fatigue and Rutting performances of the study mixes are shown respectively in Chapter V and VI. Through these laboratory test data the CalME I-R models parameters for Master Curve, fatigue damage and permanent shear strain were evaluated. Lastly, in Chapter VII, the results of the asphalt pavement structures simulations with different surface layers were reported. For each pavement structure, the total surface cracking, the total rutting, the fatigue damage and the rutting depth in each bound layer were analyzed.

Predictors of vaginal practices for sex and hygiene in KwaZulu-Natal, South Africa: findings of a household survey and qualitative inquiry

Vaginal practices in sub-Saharan Africa may increase HIV transmission and have important implications for development of microbicides and future HIV prevention technologies. It remains unclear which women undertake vaginal practices and what factors predict prevalence, practice type and choice of products. Using cross-sectional data from mixed research methods, we identify factors associated with vaginal practices among women in KwaZulu-Natal, South Africa. Data were gathered through focus group discussions, in-depth and key-informant interviews, followed by a province-wide, multi-stage cluster household survey, using structured questionnaires in face-to-face interviews with 867 women. This paper details six types of vaginal practices, which--despite their individual distinctiveness and diverse motivations--may be clustered into two broad groups: those undertaken for purposes of 'hygiene' (genital washing, douching and application) and those for 'sexual motivations' (application, insertion, ingestion and incisions). Multivariate analysis found significant associations between 'hygiene' practices and media access, religiosity and transactional sex. 'Sexual' practices were associated with partner concurrency, religiosity and use of injectable hormonal contraceptives. Future interventions relating to vaginal practices as well as microbicides need to reflect this characterisation of practices as sexual- and/or hygiene-related.

Comparison of partial least squares regression and principal component regression for pelvic shape prediction.

This paper studied two different regression techniques for pelvic shape prediction, i.e., the partial least square regression (PLSR) and the principal component regression (PCR). Three different predictors such as surface landmarks, morphological parameters, or surface models of neighboring structures were used in a cross-validation study to predict the pelvic shape. Results obtained from applying these two different regression techniques were compared to the population mean model. In almost all the prediction experiments, both regression techniques unanimously generated better results than the population mean model, while the difference on prediction accuracy between these two regression methods is not statistically significant (α=0.01).

A comparison of Markov and generalized linear models of hospital mortality

This paper reports a comparison of three modeling strategies for the analysis of hospital mortality in a sample of general medicine inpatients in a Department of Veterans Affairs medical center. Logistic regression, a Markov chain model, and longitudinal logistic regression were evaluated on predictive performance as measured by the c-index and on accuracy of expected numbers of deaths compared to observed. The logistic regression used patient information collected at admission; the Markov model was comprised of two absorbing states for discharge and death and three transient states reflecting increasing severity of illness as measured by laboratory data collected during the hospital stay; longitudinal regression employed Generalized Estimating Equations (GEE) to model covariance structure for the repeated binary outcome. Results showed that the logistic regression predicted hospital mortality as well as the alternative methods but was limited in scope of application. The Markov chain provides insights into how day to day changes of illness severity lead to discharge or death. The longitudinal logistic regression showed that increasing illness trajectory is associated with hospital mortality. The conclusion is reached that for standard applications in modeling hospital mortality, logistic regression is adequate, but for new challenges facing health services research today, alternative methods are equally predictive, practical, and can provide new insights. ^

Discrete (MD) and continuum (DD) simulations of void growth in single crystals.

Void growth in ductile materials is an important problem from the fundamental and technological viewpoint. Most of the models developed to quantify and understand the void growth process did not take into account two important factors: the anisotropic nature of plastic flow in single crystals and the size effects that appear when plastic flow is confined into very small regions.