Search

This paper presents a comprehensive investigation into the performance of multi-asset funds investing internationally. Based on a custom-built conditional multi-factor model, which includes several bond and equity-related factors, along with time-varying betas and alphas, we show that funds in our dataset significantly underperform from 2004 to 2021. This evidence holds even on a before-fee basis, with funds investing predominantly in bonds exhibiting significantly higher alphas than funds investing predominantly in equities. Since multi-asset funds may better hedge against market downturns than equity or bond funds, given their higher asset class diversification, we also evaluate performance for crisis and non-crisis phases separately. The results show that, during market crises, international multi-asset funds perform significantly better than in non-crisis times, achieving neutral performance. However, while funds that favour bond investments perform similarly across different market phases, funds that prioritize equity investments perform significantly better during crises. Consequently, multi-asset funds with higher bond holdings can be a better option during non-crisis periods, while funds with higher equity holdings should be preferred during market downturns. This somewhat puzzling finding seems to be related to funds' cash holdings. By providing a better understanding of the asset allocation decisions of international mutual fund managers, as well as of their impact on fund performance, this work has meaningful implications for investors. [ABSTRACT FROM AUTHOR]

Self-supervised learning (SSL) is a powerful technique for learning from unlabeled data. By learning to remain invariant to applied data augmentations, methods such as SimCLR and MoCo can reach quality on par with supervised approaches. However, this invariance may be detrimental for solving downstream tasks that depend on traits affected by augmentations used during pretraining, such as color. In this paper, we propose to foster sensitivity to such characteristics in the representation space by modifying the projector network, a common component of self-supervised architectures. Specifically, we supplement the projector with information about augmentations applied to images. For the projector to take advantage of this auxiliary conditioning when solving the SSL task, the feature extractor learns to preserve the augmentation information in its representations. Our approach, coined C onditional A ugmentation-aware S elf- s upervised Le arning (CASSLE), is directly applicable to typical joint-embedding SSL methods regardless of their objective functions. Moreover, it does not require major changes in the network architecture or prior knowledge of downstream tasks. In addition to an analysis of sensitivity towards different data augmentations, we conduct a series of experiments, which show that CASSLE improves over various SSL methods, reaching state-of-the-art performance in multiple downstream tasks. 1 1 We make our codebase available at https://github.com/gmum/CASSLE. 2 2 The supplementary material of this paper was shared separately due to page count limit (see Appendix A). The entire paper, along with the supplementary material, can also be accessed at https://arxiv.org/abs/2306.06082. 3 3 A short version of this paper appeared at the NeurIPS 2023 Workshop: Self-Supervised Learning-Theory and Practice (https://sslneurips23.github.io). [ABSTRACT FROM AUTHOR]

This study analyzes the relationship between homicides due to the armed conflict and educational efficiency in 912 municipalities in Colombia between 2014 and 2018, through the Saber 11 standardized test and school pass rates. The main objective is to calculate the potential loss of efficiency in the municipalities due to the armed conflict; to do this, a conditional robust non-parametric approach is applied, incorporating armed conflict as an environmental variable. Results show, primarily, that armed conflict deaths in the municipality are associated with lower educational efficiency; additionally, conflict in both local and neighboring municipalities is found to be significantly correlated with level of educational inefficiency. • Conflict-education link explored in Colombian municipalities. • Post-2016 peace agreement period studied. • First to measure armed conflict's relation on educational efficiency. • Novel use of standardized tests for conflict analysis. • Spatial contagion models incorporate neighboring conflict intensity [ABSTRACT FROM AUTHOR]

Motivated by the analysis of behavioral data taken from an economic experiment based on the Hawk-and-Dove game, this article describes a multilevel hidden Markov model, that includes covariates, autoregression, and endogenous initial conditions under a unified framework. The data at hand are affected by multiple sources of latent heterogeneity, due to multilevel unobserved factors that operate in conjunction with observed covariates at all the levels of the data hierarchy. We fit a multilevel logistic regression model for repeated measurements of player behaviors, nested within groups of interacting players. The model integrates discrete random effects at the group level and Markovian sequences of discrete random effects at the player level. Parameters are estimated by a computationally feasible expectation-maximization algorithm. We model the probability of playing the Hawk strategy, which implies fighting aggressively for controlling an asset, and test the role played by initial possession, property, and other player-specific characteristics in driving hawkish behaviors. The results from our study suggest that crucial factors in determining hawkish behavior are both the way possession is achieved – which depends on our treatment manipulation- and possession itself. Furthermore, a clear time-dependence is observed in the data at the player level as accounted for by the Markovian random effects. [ABSTRACT FROM AUTHOR]

This study investigates the Market Timing Ability (MTA) of large-cap equity fund managers in India. The extensions of Treynor and Mazuy (TM) model and Henriksson and Merton (HM) model have been used by adding six additional factors related to the public information, 91-days Treasury bill’s yield, the dividend yield on CNX 500 index, term structure of interest rates, the price-to-earnings ratio, yield from foreign exchange rates changes, and growth rate in gold prices. The extended models are termed, conditional models. This study has used time-series data of large-cap equity funds. The results of the conditional and unconditional versions of TM and HM models reveal that the large-cap equity funds as a whole do not possess significant MTA, even though a considerable percentage of the funds under each of the models show significant positive MTA. The study also highlights that the inclusion of the public information variables reconstitutes the impact of the market timing factor and other beta estimates in the model.

This study investigates the Market Timing Ability (MTA) of large-cap equity fund managers in India. The extensions of Treynor and Mazuy (TM) model and Henriksson and Merton (HM) model have been used by adding six additional factors related to the public information, 91-days Treasury bill's yield, the dividend yield on CNX 500 index, term structure of interest rates, the price-to-earnings ratio, yield from foreign exchange rates changes, and growth rate in gold prices. The extended models are termed, conditional models. This study has used time-series data of large-cap equity funds. The results of the conditional and unconditional versions of TM and HM models reveal that the large-cap equity funds as a whole do not possess significant MTA, even though a considerable percentage of the funds under each of the models show significant positive MTA. The study also highlights that the inclusion of the public information variables reconstitutes the impact of the market timing factor and other beta estimates in the model. [ABSTRACT FROM AUTHOR]

This paper provides new evidence on the appropriateness of the Fama-French five-factor model to evaluate international equity funds' performance. After extending this model to a conditional framework by allowing for time-varying risk and performance, the results show that funds underperformed during the 2000–2017 period. Funds investing globally and in Europe tend to invest in aggressive firms, but only the latter are exposed to firms with weak profitability. We thus conclude that although both investment and profitability factors are significant in explaining fund returns, the investment factor plays a more relevant role irrespective of the funds' geographical area of investment. [ABSTRACT FROM AUTHOR]

In this article, we deal with sparse high-dimensional multivariate regression models. The models distinguish themselves from ordinary multivariate regression models in two aspects: (1) the dimension of the response vector and the number of covariates diverge to infinity; (2) the nonzero entries of the coefficient matrix and the precision matrix are sparse. We develop a two-stage sequential conditional selection (TSCS) approach to the identification and estimation of the nonzeros of the coefficient matrix and the precision matrix. It is established that the TSCS is selection consistent for the identification of the nonzeros of both the coefficient matrix and the precision matrix. Simulation studies are carried out to compare TSCS with the existing state-of-the-art methods, which demonstrates that the TSCS approach outperforms the existing methods. As an illustration, the TSCS approach is also applied to a real dataset. [ABSTRACT FROM AUTHOR]

Recurrent events arise when an event occurs many times for a subject. Many models have been developed to analyze these kind of data: the Andersen-Gill's model is one of them as well as the Prentice-William and the Peterson's model, the Wei Lee and Weissfeld's model, or even frailty models, all assuming an independent and noninformative censoring. However, in practice, these assumptions may be violated by the existence of a terminal event that permanently stops the recurrent process (eg, death). Indeed, a patient who experiences an early terminal event is more likely to have a lower number of recurrent events than a patient who experiences a terminal event later. Thus, ignoring terminal events in the analysis may lead to biased results. Many methods have been developed to handle terminal events. In this paper, we describe the existing methods classifying into conditional or marginal methods and compare them in a simulation study to highlight bias in results if an inappropriate method is used, when recurrent events and terminal event are correlated. In addition, we apply the different models on a real dataset to show how results should be interpreted. Finally, we provide recommendations for choosing the appropriate method for analyzing recurrent events in the presence of a terminal event. [ABSTRACT FROM AUTHOR]

• This manuscript presents a thorough review of the state-of-the-art on multi-organ analysis in medical imaging. • More than 300 papers reviewed, discussed, and categorized methodologically and anatomically. • For the first time, this paper proposes a methodology-based classification of the different techniques available for the analysis of multi-organ anatomical complex, from the simple global modelling, to the more sophisticated sequential and multi-resolution techniques. • The different methodologies for multi-organ analysis are classified using the following categorization: global and individual models, coupled deformable models, multi-level models, sequential models, atlas-based models, machine-learning models, graphical models, and articulated models. • The manuscripts also reflects on the trends and challenges of multi-organ analysis, the peculiarities of each anatomical region, and its impact on the future of healthcare The medical image analysis field has traditionally been focused on the development of organ-, and disease-specific methods. Recently, the interest in the development of more comprehensive computational anatomical models has grown, leading to the creation of multi-organ models. Multi-organ approaches, unlike traditional organ-specific strategies, incorporate inter-organ relations into the model, thus leading to a more accurate representation of the complex human anatomy. Inter-organ relations are not only spatial, but also functional and physiological. Over the years, the strategies proposed to efficiently model multi-organ structures have evolved from the simple global modeling, to more sophisticated approaches such as sequential, hierarchical, or machine learning-based models. In this paper, we present a review of the state of the art on multi-organ analysis and associated computation anatomy methodology. The manuscript follows a methodology-based classification of the different techniques available for the analysis of multi-organs and multi-anatomical structures, from techniques using point distribution models to the most recent deep learning-based approaches. With more than 300 papers included in this review, we reflect on the trends and challenges of the field of computational anatomy, the particularities of each anatomical region, and the potential of multi-organ analysis to increase the impact of medical imaging applications on the future of healthcare. Image, graphical abstract [ABSTRACT FROM AUTHOR]

Copyright of Acta Automatica Sinica is the property of Chinese Academy of Sciences, Institute of Automation and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Canadian Journal of Administrative Sciences (John Wiley & Sons, Inc.) is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Canadian Journal of Administrative Sciences (John Wiley & Sons, Inc.) is the property of Wiley-Blackwell and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Copyright of Spanish Journal of Finance & Accounting / Revista Espanola de Financiacion y Contabilidad is the property of Routledge and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

En este artículo de investigación se desarrolla un análisis de las propiedades de los modelos multinomiales a través de distintos procesos de simulación. Lo anterior se hizo asumiendo tanto el cumplimiento de los supuestos subyacentes de los mecanismos de estimación como el incumplimiento de los mismos. Igualmente se analizó el comportamiento de los estimadores bajo diferentes escenarios de tamaño muestral. Se encontró que bajo un modelo correctamente especificado y tamaños muestrales superiores a 200 observaciones se cumplen las propiedades de insesgadez y consistencia, mientras que la incorrecta especificación de la distribución del proceso lleva a estimaciones sesgadas e inconsistentes; de igual forma se encontró que en tamaños muestrales pequeños y bajo modelos condicionales se pierden las propiedades que una buena especificación del proceso suele generar, y se halla aún más inestabilidad cuando la estimación es llevada a cabo con la metodología Probit.This paper develops an analysis of Multinomial models through simulation; this was done under correct and incorrect assumptions on the data generating process. Also, it was analyzed the performance of the models under different sample sizes. It was found that a correct specified model with samples of 200 or more observations achieves estimators which are unbiased and consistent, while incorrect assumptions about the data generating process causes biased and inconsistent estimators. On the other hand, conditional models with small sample sizes imply bad statistical properties, especially when Probit models are estimated.

Polytomous responses modeling, particularly ordinal ones, has been the subject of increasing interest in recent years. It has been gaining ground in research on quality of life and health status indicators, among other topics. Its use ranges from cross-sectional studies, which assume independence between observations, to longitudinal studies, in which two or more responses from the same individual are observed over time. There are several methods proposed in the literature to model data of this nature in crosssectional studies, and in practice the most common is the proportional odds model. There is also the partial proportional odds model, which generalizes the proportional odds model and allows for non-proportionality in a subset of covariates. In longitudinal studies, the usual models - marginal and generalized linear mixed (conditional) models - for analysis of correlated data can also be used to model polytomous responses. In this paper, models for ordinal polytomous responses are discussed from an applied perspective of conditional and marginal structures. The specification and interpretation of the models are illustrated and discussed by analyzing two real data sets. [ABSTRACT FROM AUTHOR]

Milton Friedman and Anna Schwartz constructed an important macroeconomic dataset for the United Kingdom that spans 1878-1970. Numerous authors have modeled the demand for broad money on that dataset. Model selection is central to assessing the merits of the resulting empirical models, so the current paper re-evaluates that issue with computer-automated model selection. Some models are robust to the model selection path, as characterized through variations in target size, pre-search testing, fixity of regressors, impulse indicator saturation, representation of the general model, and choice of dependent variable. Model improvement is also feasible, with historically interpretable nonlinearities and structural breaks. [ABSTRACT FROM AUTHOR]

Background: By 2020, the global health community aims to control and eliminate human helminthiases, including schistosomiasis in selected African countries, principally by preventive chemotherapy (PCT) through mass drug administration (MDA) of anthelminthics. Quantitative monitoring of anthelminthic responses is crucial for promptly detecting changes in efficacy, potentially indicative of emerging drug resistance. Statistical models offer a powerful means to delineate and compare efficacy among individuals, among groups of individuals and among populations. Methods: We illustrate a variety of statistical frameworks that offer different levels of inference by analysing data from nine previous studies on egg counts collected from African children before and after administration of praziquantel. Results: We quantify responses to praziquantel as egg reduction rates (ERRs), using different frameworks to estimate ERRs among population strata, as average responses, and within strata, as individual responses. We compare our modelbased average ERRs to corresponding model-free estimates, using as reference the World Health Organization (WHO) 90 % threshold of optimal efficacy. We estimate distributions of individual responses and summarize the variation among these responses as the fraction of ERRs falling below the WHO threshold. Conclusions: Generic models for evaluating responses to anthelminthics deepen our understanding of variation among populations, sub-populations and individuals. We discuss the future application of statistical modelling approaches for monitoring and evaluation of PCT programmes targeting human helminthiases in the context of the WHO 2020 control and elimination goals. [ABSTRACT FROM AUTHOR]

Detailed segmentation of the vertebrae is an important pre-requisite in various applications of image-based spine assessment, surgery and biomechanical modeling. In particular, accurate segmentation of the processes is required for image-guided interventions, for example for optimal placement of bone grafts between the transverse processes. Furthermore, the geometry of the processes is now required in musculoskeletal models due to their interaction with the muscles and ligaments. In this paper, we present a new method for detailed segmentation of both the vertebral bodies and processes based on statistical shape decomposition and conditional models. The proposed technique is specifically developed with the aim to handle the complex geometry of the processes and the large variability between individuals. The key technical novelty in this work is the introduction of a part-based statistical decomposition of the vertebrae, such that the complexity of the subparts is effectively reduced, and model specificity is increased. Subsequently, in order to maintain the statistical and anatomic coherence of the ensemble, conditional models are used to model the statistical inter-relationships between the different subparts. For shape reconstruction and segmentation, a robust model fitting procedure is used to exclude improbable inter-part relationships in the estimation of the shape parameters. Segmentation results based on a dataset of 30 healthy CT scans and a dataset of 10 pathological scans show a point-to-surface error improvement of 20% and 17% respectively, and the potential of the proposed technique for detailed vertebral modeling. [ABSTRACT FROM AUTHOR]

Longitudinal experiments often involve multiple outcomes measured repeatedly within a set of study participants. While many questions can be answered by modeling the various outcomes separately, some questions can only be answered in a joint analysis of all of them. In this article, we will present a review of the many approaches proposed in the statistical literature. Four main model families will be presented, discussed and compared. Focus will be on presenting advantages and disadvantages of the different models rather than on the mathematical or computational details. [ABSTRACT FROM PUBLISHER]

Resource selection functions (RSF) are often developed using satellite (ARGOS) or Global Positioning System (GPS) telemetry datasets, which provide a large amount of highly correlated data. We discuss and compare the use of generalized linear mixed-effects models (GLMM) and generalized estimating equations (GEE) for using this type of data to develop RSFs. GLMMs directly model differences among caribou, while GEEs depend on an adjustment of the standard error to compensate for correlation of data points within individuals. Empirical standard errors, rather than model-based standard errors, must be used with either GLMMs or GEEs when developing RSFs. There are several important differences between these approaches; in particular, GLMMs are best for producing parameter estimates that predict how management might influence individuals, while GEEs are best for predicting how management might influence populations. As the interpretation, value, and statistical significance of both types of parameter estimates differ, it is important that users select the appropriate analytical method. We also outline the use of k-fold cross validation to assess fit of these models. Both GLMMs and GEEs hold promise for developing RSFs as long as they are used appropriately.

We group approaches to modeling correlated binary data according to data recorded cross-sectionally as opposed to data recorded longitudinally; according to models that are population-averaged as opposed to subject-specific; and according to data with time-dependent covariates as opposed to time-independent covariates. Standard logistic regression models are appropriate for cross-sectional data. However, for longitudinal data, methods such as generalized estimating equations (GEE) and generalized method of moments (GMM) are commonly used to fit population-averaged models, while random-effects models such as generalized linear mixed models (GLMM) are used to fit subject-specific models. Some of these methods account for time-dependence in covariates while others do not. This paper addressed these approaches with an illustration using a Medicare dataset as it relates to rehospitalization. In particular, we compared results from standard logistic models, GEE models, GMM models, and random-effects models by analyzing a binary outcome for four successive hospitalizations. We found that these procedures address differently the correlation among responses and the feedback from response to covariate. We found marginal GMM logistic regression models to be more appropriate when covariates are classified as time-dependent in comparison to GEE models. We also found conditional random-intercept models with time-dependent covariates decomposed into components to be more appropriate when time-dependent covariates are present in comparison to ordinary random-effects models. We used the SAS procedures GLIMMIX, NLMIXED, IML, GENMOD, and LOGISTIC to analyze the illustrative dataset, as well as unique programs written using the R language. [ABSTRACT FROM AUTHOR]

Summary: Conditional mouse models based on the Cre-recombinase (Cre)-loxP method are a powerful tool for determining the spatial and temporal function of genes in neuroscience research. The Emx1-Cre conditional model is designed to drive Cre expression in a predominantly excitatory neuron specific manner and the Dlx5/6-Cre mouse expresses Cre predominantly in cortical inhibitory neurons. The mouse models expressing the Cre transgene are healthy, active and have no overt behavioural or brain histological phenotypes. Subcutaneous pentylenetetrazol (scPTZ) is a proconvulsant frequently used to probe neuronal network excitability. In both the Emx1-Cre and Dlx5/6-Cre conditional mouse models the latency to scPTZ-induced seizures was significantly shorter than for their wild-type littermates. This shows that mouse models carrying the Cre transgene alone can have significant behavioural phenotypes. This may act as a confound to the interpretation of data obtained from crosses with loxP-flanked targets especially in the context of epilepsy phenotypes. These data highlight that appropriate control experiments that compare wild-type mice to those that carry the cre-transgene but not the loxP-flanked target are essential when using this method. [Copyright &y& Elsevier]

ABSTRACT This paper investigates the style and performance of US and European global socially responsible funds. Several specifications of the return-generating process are applied as well as their corresponding conditional versions. Most European global socially responsible funds do not show significant performance differences in relation to both conventional and socially responsible benchmarks. US funds and Austrian funds show evidence of underperformance. By applying conditional models, we find evidence of time-varying betas but not of time-varying alphas. With respect to investment style, we find evidence that socially responsible funds are strongly exposed to small cap and growth stocks. Although these results are consistent with previous studies, they uncover some misclassification issues in these funds. Finally, we also document a significant home bias for global socially responsible funds. Copyright © 2011 John Wiley & Sons, Ltd. [ABSTRACT FROM AUTHOR]

The presence of one or more covariates that perfectly or almost perfectly predict the outcome of interest (which is referred to as complete or quasi-complete separation, the latter denoting the case when such perfect prediction occurs only for a subset of observations in the data) has been extensively studied in the last four decades. Since 1984, when Albert and Anderson (1984) differentiated between complete and quasi-complete separation, several authors have studied this phenomenon and tried to provide answers or ways of identifying the problem (Lesaffre and Albert, 1989; Firth, 1993; Christmann and Rousseeuw, 2001; Rousseeuw and Christmann, 2003; Allison, 2004; Zorn, 2005; Heinze, 2006). From an estimation perspective, separation leads to infinite coefficients and standard errors, which makes the algorithm collapse or give inappropriate results. As a practical matter, separation forces the analyst to choose from a number of problematic alternatives for dealing with the problem, and in the past the elimination of such problematic variables were common practice to deal with such situations. In the last decade, solutions using penalized likelihood have been proposed, but always dealing with independent binary data. Here we will propose a Bayesian solution to the problem when we deal with clustered binary data using informative priors that are supported by the data and compare it with an alternative procedure proposed by Gelman et al. (2008). [ABSTRACT FROM AUTHOR]

Resource selection functions (RSF) are often developed using satellite (ARGOS) or Global Positioning System (GPS) telemetry datasets, which provide a large amount of highly correlated data. We discuss and compare the use of generalized linear mixed-effects models (GLMM) and generalized estimating equations (GEE) for using this type of data to develop RSFs. GLMMs directly model differences among caribou, while GEEs depend on an adjustment of the standard error to compensate for correlation of data points within individuals. Empirical standard errors, rather than model-based standard errors, must be used with either GLMMs or GEEs when developing RSFs. There are several important differences between these approaches; in particular, GLMMs are best for producing parameter estimates that predict how management might influence individuals, while GEEs are best for predicting how management might influence populations. As the interpretation, value, and statistical significance of both types of parameter estimates differ, it is important that users select the appropriate analytical method. We also outline the use of k-fold cross validation to assess fit of these models. Both GLMMs and GEEs hold promise for developing RSFs as long as they are used appropriately. [ABSTRACT FROM AUTHOR]

In this research paper, we made an analysis of Multinomial models through simulation; this was done under correct and incorrect assumptions on the data generating process. Also, it was analyzed the performance of the models under different sample sizes. It was found that a correct specified model with samples of 200 or more observations achieves estimators which are unbiased and consistent, while incorrect assumptions about the data generating process causes biased and inconsistent estimators. On the other hand, conditional models with small sample sizes imply bad statistical properties, especially when Probit models are estimated. [ABSTRACT FROM AUTHOR]

The growth of information available to learning systems and the increasing complexity of learning tasks determine the need for devising algorithms that scale well with respect to all learning parameters. In the context of supervised sequential learning, the Viterbi algorithm plays a fundamental role, by allowing the evaluation of the best (most probable) sequence of labels with a time complexity linear in the number of time events, and quadratic in the number of labels. In this paper we propose CarpeDiem, a novel algorithm allowing the evaluation of the best possible sequence of labels with a sub-quadratic time complexity. We provide theoretical grounding together with solid empirical results supporting two chief facts. CarpeDiem always finds the optimal solution requiring, in most cases, only a small fraction of the time taken by the Viterbi algorithm; meantime, CarpeDiem is never asymptotically worse than the Viterbi algorithm, thus confirming it as a sound replacement. [ABSTRACT FROM AUTHOR]

This paper examines the performance of Portuguese equity funds investing in the domestic and in the European Union market, using several unconditional and conditional multi-factor models. In terms of overall performance, we find that National funds are neutral performers, while European Union funds under-perform the market significantly. These results do not seem to be a consequence of management fees. Overall, our findings are supportive of the robustness of conditional multi-factor models. In fact, Portuguese equity funds seem to be relatively more exposed to smallcaps and more value-oriented. Also, they present strong evidence of time-varying betas and, in the case of the European Union funds, of time-varying alphas too. Finally, in terms of market timing, our tests suggest that mutual fund managers in our sample do not exhibit any market timing abilities. Nevertheless, we find some evidence of timevarying conditional market timing abilities but only at the individual fund level. [ABSTRACT FROM AUTHOR]

This paper presents the first comprehensive study of the performance and market timing ability of the equity funds that comprise the Hong Kong Mandatory Provident Funds (MPF) scheme. In general, our results suggest that US equity funds consistently underperform relative to the market, while the other fund groups consistently outperform the market. The stock-selection ability of MPF constituent equity funds in times of changing economic condition is also investigated. The evidence is consistent with previous studies, which suggest that the conditional models decrease the individual fund traditional alpha measure. The market timing models of Treynor–Mazuy and Henriksson–Merton provide evidence of superior market timing ability. [ABSTRACT FROM AUTHOR]

Abstract: Health care quality improvement interventions are often evaluated in randomized trials in which individual physicians serve as the unit of randomization. These cluster randomized trials present a unique data structure that consists of many clusters of highly variable size. The appropriate method of analysis for these trials is unknown. We conducted a simulation study to compare several methods for analyzing data which were generated to replicate the structure of our motivating example. We varied the treatment effect size and the distributional assumptions about the random effect. Simulation was used to estimate power, coverage, bias, and mean squared error of full maximum likelihood estimation (MLE), approximate MLE using penalized quasi-likelihood (PQL), generalized estimating equations (GEE), group-bootstrapped logistic regression, and a clustered permutation test. Across all conditions tested, GEE and full MLE performed comparably. Bootstrapped methods were less powerful and had higher mean squared error under conditions of variable cluster size. PQL yielded biased results. The permutation test preserved Type I error rates, but had less power than the other methods considered. [Copyright &y& Elsevier]

Previous research suggests that one field with a strong yet unsatisfied need for automatically extracting instances of various entity classes from texts is the analysis of socio-technical systems (Feldstein in Media in Transition MiT5, ; Hampe et al. in Netzwerkanalyse und Netzwerktheorie, ; Weil et al. in Proceedings of the 2006 Command and Control Research and Technology Symposium, ; Diesner and Carley in XXV Sunbelt Social Network Conference, ). Traditional as well as non-traditional and customized sets of entity classes and the relationships between them are often specified in ontologies or taxonomies. We present a Conditional Random Fields (CRF)-based approach to distilling a set of entities that are defined in an ontology originating from organization science. CRF, a supervised sequential machine learning technique, facilitates the derivation of relational data from corpora by locating and classifying instances of various entity classes. The classified entities can be used as nodes for the construction of socio-technical networks. We find the outcome sufficiently accurate (82.7 percent accuracy of locating and classifying entities) for future application in the described problem domain. We propose using the presented methodology as a crucial step in the process of advanced modeling and analysis of complex and dynamic networks. [ABSTRACT FROM AUTHOR]

Abstract: We describe algorithms for recognizing human motion in monocular video sequences, based on discriminative conditional random fields (CRFs) and maximum entropy Markov models (MEMMs). Existing approaches to this problem typically use generative structures like the hidden Markov model (HMM). Therefore, they have to make simplifying, often unrealistic assumptions on the conditional independence of observations given the motion class labels and cannot accommodate rich overlapping features of the observation or long-term contextual dependencies among observations at multiple timesteps. This makes them prone to myopic failures in recognizing many human motions, because even the transition between simple human activities naturally has temporal segments of ambiguity and overlap. The correct interpretation of these sequences requires more holistic, contextual decisions, where the estimate of an activity at a particular timestep could be constrained by longer windows of observations, prior and even posterior to that timestep. This would not be computationally feasible with a HMM which requires the enumeration of a number of observation sequences exponential in the size of the context window. In this work we follow a different philosophy: instead of restrictively modeling the complex image generation process – the observation, we work with models that can unrestrictedly take it as an input, hence condition on it. Conditional models like the proposed CRFs seamlessly represent contextual dependencies and have computationally attractive properties: they support efficient, exact recognition using dynamic programming, and their parameters can be learned using convex optimization. We introduce conditional graphical models as complementary tools for human motion recognition and present an extensive set of experiments that show not only how these can successfully classify diverse human activities like walking, jumping, running, picking or dancing, but also how they can discriminate among subtle motion styles like normal walks and wander walks. [Copyright &y& Elsevier]

Many applications of object recognition in the presence of pose uncertainty rely on statistical models--conditioned on pose--for observations. The image statistics of three-dimensional (3-D) objects are often assumed to belong to a family of distributions with unknown model parameters that vary with one or more continuous-valued pose parameters. Many methods for statistical model assessment, for example the tests of Kolmogorov-Smirnov and K. Pearson, require that all model parameters be fully specified or that sample sizes be large. Assessing pose-dependent models from a finite number of observations over a variety of poses can violate these requirements. However, a large number of small samples, corresponding to unique combinations of object, pose, and pixel location, are often available. We develop methods for model testing which assume a large number of small samples and apply them to the comparison of three models for synthetic aperture radar images of 3-D objects with varying pose. Each model is directly Sated to the Gaussian distribution and is assessed both in terms of goodness-of-fit and underlying model assumptions, such as independence, known mean, and homoscedasticity. Test results are presented in terms of the functional relationship between a given significance level and the percentage of samples that wold fall a test at that level. [ABSTRACT FROM AUTHOR]

In this paper we evaluate the performance of European bond funds using unconditional and conditional models. As conditioning information we use variables that we find to be useful in predicting bond returns in the European market. The results show that, in general, bond funds are not able to outperform passive strategies. These findings are robust to whatever model (unconditional versus conditional and single versus multi-index) we use. The multi-index model seems to add some explanatory power in relation to the single-index model. Furthermore, when we incorporate the predetermined information variables, we can observe a slight tendency towards better performance. This evidence is consistent with previous studies on stock funds and comes in support of the argument that conditional models might allow for a better assessment of performance. However, our results suggest that the impact of additional risk factors seems to be greater than the impact of incorporating predetermined information variables. [ABSTRACT FROM AUTHOR]

Structural analysis is applied to exploit sparsity in the solving of a system of equations [Direct Methods for Sparse Matrices. Monographs on Numerical Analysis (1989)]. Zaher [Conditional modeling (1995)] studied the issues involved in the structural analysis of conditional models and presented a methodology to ensure consistency in a conditional model, the complexity of such an analysis being combinatorial. In that work Zaher considered only cases in which the number of variables and equations of all the alternatives in a conditional model are the same. In this paper, an extension to Zaher''s consistency analysis is presented. This extension allows the consistency analysis to be applied to the conditional models in which the number of variables and equations for each of the alternatives may not be the same. Also, we show how, by taking advantage of the structure of the problem, it is sometimes possible to reduce the combinatorial effort required by such an analysis. In particular, the cases of the existence of repeated structures and common incidence pattern among alternatives are discussed. [Copyright &y& Elsevier]

Interior point methods have recently become an interesting alternative in a number of numerical applications. In particular, their performance in the solution of problems involving complementarity equations has been the subject of extensive research and their efficacy is well documented. In this paper, following a brief description of the fundamentals of interior point methods and the globally convergent framework proposed by Wang et al. (Mathematical Programming 74 (1996) 159), we show how we can apply such an algorithm for solving the complementarity representation of a conditional model (Industrial Engineering and Chemical Research 38 (1999) 519), where such models involve sums of complementary products being zero. Furthermore, we modified Wang''s algorithm in order to apply a high order strategy designed to improve convergence (SIAM Journal of Optimization 2 (1992) 575). We then use the proposed approach to solve some conditional models encountered in the field of chemical engineering. This technique has been incorporated into the ascend modeling environment with the implementation of the solver ipslv. [Copyright &y& Elsevier]