Model Comparison Tests Research Articles

A comparison of factor models in China

We evaluate the performance of eleven asset pricing models in the Chinese A-share market using a variety of test portfolios and statistical methodologies. To compile the test portfolios, we construct 105 anomalies and use the 23 significant anomalies as test assets for model comparison. The results indicate that, in time-series test and anomaly explanations, the Hou et al. (2019) five-factor q model demonstrates the best overall performance. The pairwise cross-sectional R2 tests and multiple model comparison tests further affirm that the Hou et al. (2019) five-factor q model, the Fama and French (2018) six-factor (FF6) model, and the Kelly et al. (2019) five-factor Instrumented Principal Component Analysis (IPCA5) model are the top performers. Notably, the performance of the five-factor q model remains robust across various experimental designs.

The impact of various substrates, ceramic shades, and brands on the ultimate color and masking capacity of highly translucent monolithic zirconia: an in vitro study.

This study aimed to examine the impact of substrates, ceramic shades, and brands on the color and masking ability of highly translucent monolithic zirconia (HTMZ) using CIELab and CIEΔE2000 metrics. A total of 156 1-mm thick HTMZ disks in shades A1, A2, and A3 were produced using Dental Direkt and Kerox zirconia brands. Four 3-mm thick substrates (nickel-chromium alloy, non-precious gold alloy (NPG), zirconia shade A2, and resin composite shade A2) were prepared. HTMZ disks were overlaid on these substrates, and color measurements were taken with a spectrophotometer. Color differences (ΔE) were analyzed using CIELab and CIEΔE2000 formulas. The influence of brand, shade, substrate, and their interactions on ΔE values was assessed with a General Linear Model (GLM) and LSD pairwise comparison test. Spearman's correlation test examined the relationship between CIELab and CIEΔE2000 values. Results indicated that ΔEab was significantly influenced by substrate type and shade, while ΔE2000 was also affected by the ceramic brand. Mean color differences across ceramic-substrate groups were within clinically acceptable and perceptible ranges (clinically perceptible: ∆Eab ≥ 1.3 and ∆E2000 ≥ 0.8; clinically acceptable: 0.8 < ∆E2000 ≤ 1.8 and 1.3 < ∆Eab ≤ 2.7), except for NPG, which had ΔE values exceeding the perceptible range (ΔE2000: 1.1 ± 0.11 to 1.8 ± 0.31; ΔEab: 1.61 ± 0.15 to 2.16 ± 0.36). A significant correlation (r = 0.974, P < 0.001) was found between ΔEab and ΔE2000. Various ceramic brands and shades led to notable ΔE variations, yet average color differences within all ceramic-substrate groups remained clinically acceptable. Both ΔEab and ΔE2000 were reliable methods with a strong correlation for measuring color differences.

A bootstrap model comparison test for identifying genes with context-specific patterns of genetic regulation.

Understanding how genetic variation affects gene expression is essential for a complete picture of the functional pathways that give rise to complex traits. Although numerous studies have established that many genes are differentially expressed in distinct human tissues and cell types, no tools exist for identifying the genes whose expression is differentially regulated. Here we introduce DRAB (differential regulation analysis by bootstrapping), a gene-based method for testing whether patterns of genetic regulation are significantly different between tissues or other biological contexts. DRAB first leverages the elastic net to learn context-specific models of local genetic regulation and then applies a novel bootstrap-based model comparison test to check their equivalency. Unlike previous model comparison tests, our proposed approach can determine whether population-level models have equal predictive performance by accounting for the variability of feature selection and model training. We validated DRAB on mRNA expression data from a variety of human tissues in the Genotype-Tissue Expression (GTEx) Project. DRAB yielded biologically reasonable results and had sufficient power to detect genes with tissue-specific regulatory profiles while effectively controlling false positives. By providing a framework that facilitates the prioritization of differentially regulated genes, our study enables future discoveries on the genetic architecture of molecular phenotypes.

AN examination of linear factor models in U.K. stock returns in the presence of dynamic trading

This study uses the approach of Ferson and Siegel, Rev Financ Stud 22:2735–2758 (2009), and Ferson, Siegel and Wang, J Financ Quant Anal, forthcoming, (2024) to examine the unconditional mean–variance efficiency, in the presence of conditioning information (UMV), of ten linear factor models in U.K. stock returns. The study finds that the UMV efficiency of all the multifactor models is strongly rejected in U.K. stock returns in two different sets of test assets. This rejection is mainly driven by allowing dynamic trading in the test assets and factors. The optimal use of conditioning information also has a significant impact in relative model comparison tests. In relative model comparison tests based on UMV efficiency, the best performing model is the eight-factor model of Chib and Zeng, J Bus Econ Stat 38:771–783 (2020) model.

Structural equation modeling for the effects of family dysfunctions and communication on perceived mental health status among under/graduate students in the U.S.

Many of the college students expressed experiencing moderate or severe psychological distress. In their emergence as adults, college students encounter significant periods of drastic change. Family functions play a crucial role in determining individuals' mental health status. The first objective of this study was to investigate whether family communication mediates the relationships between family dysfunctions and perceived mental health status among under/graduate students. The second objective was to investigate whether the mediation effects of family communication on the relationships between family dysfunctions and perceived mental health status differ by gender. The current study applied Minuchin's Structural Family Theory with under/graduate students (N = 348) in Ohio. Structural Equation Modeling was conducted to examine the effects of family dysfunctions and communication on perceived mental health status among under/graduate students. Disengaged family function significantly influenced family communication, which, in turn, significantly impacted perceived mental health (mediation effect). Enmeshed family function did not show a significant relationship with family communication, whereas family communication had a significant impact on perceived mental health (no mediation effect). The model comparison test indicated that there were differences between males and females: the structural weights of the current path models were consistent with the full models, and there was a significant difference in the effect of parents' marital status on perceived mental health, with males being significantly impacted in both models. Practitioners should recognize that improving positive communication skills within the family unit can be a realistic approach to supporting the perceived mental health of under/graduate students. Colleges/universities should offer regular education programs to promote an understanding of the relationship between family communication strategies and the mental health of under/graduate students.

US efficient factors in a Bayesian model scan framework

PurposeThe author examines the impact these efficient factors have on factor model comparison tests in US returns using the Bayesian model scan approach of Chib et al. (2020), and Chib et al.(2022).Design/methodology/approachEhsani and Linnainmaa (2022) show that time-series efficient investment factors in US stock returns span and earn 40% higher Sharpe ratios than the original factors.FindingsThe author shows that the optimal asset pricing model is an eight-factor model which contains efficient versions of the market factor, value factor (HML) and long-horizon behavioral factor (FIN). The findings show that efficient factors enhance the performance of US factor model performance. The top performing asset pricing model does not change in recent data.Originality/valueThe author is the only one to examine if the efficient factors developed by Ehsani and Linnainmaa (2022) have an impact on model comparison tests in US stock returns.

Feature aggregation-based multi-relational knowledge reasoning for COPD intelligent diagnosis

The increasing prevalence of artificial intelligence-based knowledge reasoning has contributed to more accurate and efficient auxiliary diagnoses. However, a majority of the disease prediction methods concentrate on the symptoms themselves while discarding the inherent properties of symptoms and the relationships underlying them. This paper proposes a feature aggregation-based intelligent diagnosis model employing a Heterogeneous Graph Convolutional Network (GCN), termed HeteroGCN. It focuses on symptoms’ inherent properties and multiple hidden relationships among symptoms and properties. By aggregating features of nodes, it realizes effective and accurate symptom-based knowledge reasoning for disease-type prediction. The diagnosis-related information from the Electronic Medical Record (EMR) has been extracted and standardized by taking chronic obstructive pulmonary disease (COPD) as an instance. Then the presented model extracts the symptoms and their properties as nodes and the relationships underlying the nodes as edges to construct a heterogeneous graph. The adjacency matrix and feature matrix have been fused and taken as the input of this model, and then the node representations (embeddings) are generated by aggregating neighbor nodes’ information. Finally, specific disease types (syndromes) will be predicted by the generated symptom node embeddings. The results of the model comparison and parameter sensitivity test demonstrate that the presented HeteroGCN model performs best on disease-type prediction. This paper provides a novel feature aggregation-based multi-relational knowledge reasoning approach for disease type (syndrome) prediction, which holds great significance in improving disease diagnosis.

Priced risk in corporate bonds

Recent studies document strong empirical support for multifactor models that aim to explain the cross-sectional variation in corporate bond expected excess returns. We revisit these findings and provide evidence that common factor pricing in corporate bonds is exceedingly difficult to establish. Based on portfolio- and bond-level analyses, we demonstrate that previously proposed bond risk factors, with traded liquidity as the only marginal exception, do not have any incremental explanatory power over the corporate bond market factor. Consequently, this implies that the bond CAPM is not dominated by either traded- or nontraded-factor models in pairwise and multiple model comparison tests.

Anomaly Detection Algorithm for Heterogeneous Wireless Networks Based on Cascaded Convolutional Neural Networks

As the popularity of wireless networks deepens, the diversity of device types and hardware environments makes network data take on heterogeneous forms while the threat of malicious attacks from outside can prevent ordinary methods from mining information from abnormal data. In view of this, the research will be devoted to the feature processing of the anomalous data itself, and the convolutional operation of the anomalous information by the convolutional neural network (CNN). This is to extract the internal information. In the first step of the cascaded CNNs, the dimensions of the anomaly data will be processed, the anomaly data will be sorted under the concept of relevance grouping, and then the sorted results will be added to the convolution and pooling. The performance test uses three datasets with different feature capacities as the attack sources, and the results show a 13.22% improvement in information mining performance compared to the standard CNN. The extended CNN step will perform feature identification for homologous or similar network threats, with feature expansion within the convolutional layer first, and then pooling to reduce the computational cost. The test results show that when the maximum value domain of linear expansion is 2, the model has the best feature recognition performance, fluctuating around 85%; The model comparison test results show that the accuracy of the extended CNN is higher than that of the standard CNN, and the model stability is better than that of the back propagation (BP) neural network. This indicates that the cascaded CNN dual module can mine for the data itself, thus ignoring the risk unknowns, and this connected CNN has some practical significance. The proposed cascaded CNN module applies advanced neural network technology to identify internal and external risk data. The research content has important reference value for the security management of IoT systems.

A general approximation to nested Bayes factors with informed priors

A staple of Bayesian model comparison and hypothesis testing Bayes factors are often used to quantify the relative predictive performance of two rival hypotheses. The computation of Bayes factors can be challenging, however, and this has contributed to the popularity of convenient approximations such as the Bayesian information criterion (BIC). Unfortunately, these approximations can fail in the case of informed prior distributions. Here, we address this problem by outlining an approximation to informed Bayes factors for a focal parameter . The approximation is computationally simple and requires only the maximum likelihood estimate and its standard error. The approximation uses an estimated likelihood of and assumes that the posterior distribution for is unaffected by the choice of prior distribution for the nuisance parameters. The resulting Bayes factor for the null hypothesis versus the alternative hypothesis is then easily obtained using the Savage–Dickey density ratio. Three real‐data examples highlight the speed and closeness of the approximation compared with bridge sampling and Laplace's method. The proposed approximation facilitates Bayesian reanalyses of standard frequentist results, encourages application of Bayesian tests with informed priors, and alleviates the computational challenges that often frustrate both Bayesian sensitivity analyses and Bayes factor design analyses. The approximation is shown to suffer under small sample sizes and when the posterior distribution of the focal parameter is substantially influenced by the prior distributions on the nuisance parameters. The proposed methodology may also be used to approximate the posterior distribution for under .

Likelihood Ratio Search Procedure for Optimum Number of States in a Hidden Markov Manpower Model

This paper aims at obtaining the optimum number of states for a hidden Markov manpower model, which, hitherto, has been chosen arbitrarily. A search procedure that attains this optimum number after a few steps across a series of N hidden Markov manpower models is proposed. The likelihood ratio statistic is employed to conduct pairwise model comparison tests on the N hidden Markov manpower models ordered according to their level of parsimony. The illustration shows the usefulness of the procedure in choosing the right number of states for a hidden Markov manpower model to avoid wrong specification of such models. The proposed procedure can be useful in other areas of research, such as in biological, medical and social sciences, where application of hidden Markov model may require the determination of number of hidden states based on unobserved data with latent heterogeneity. The procedure has a straightforward formulation and its application in other areas requires mainly the adaptation of the model specifications for the new area’s system dynamics.

Differential effects of age and sex on medial temporal lobe networks revealed by longitudinal analyses across the adult lifespan

AbstractBackgroundThe medial temporal lobe (MTL) is extensively connected to the rest of the brain through two specific networks which are particularly affected in Alzheimer’s disease (AD): the anterior‐temporal (AT) and posterior‐medial (PM) systems. As the specific and respective effects of age and sex on the functional integrity of these networks are still largely unknown, we sought to fill this gap using longitudinal data spanning the entire adult life.MethodLongitudinal data from 215 cognitively unimpaired adults (aged 19‐85, including 113 females, followed‐up up to 47 months, 419 scans) were analysed. Seed‐based analyses were applied to resting‐state functional magnetic resonance imaging to generate connectivity maps (seeds: the perirhinal cortex for AT and the parahippocampal cortex for PM). The AT and PM networks were defined by comparing baseline maps from adults younger than 40 using a paired permutation t‐test with threshold‐free cluster enhancement (TFCE) and familywise error (FWE) correction (p&lt;.05) (Figure 1). These masks were then used to extract the mean functional connectivity (FC) for each participant. The interaction effects between network, age and sex were investigated using linear mixed models.ResultThere was a significant interaction between age and networks (p&lt;.001), with a decrease of FC with age in PM and an increase in AT (Figure 2). Sex differences were also observed within these two systems (p&lt;.001), with males exhibiting higher FC than females in AT but lower FC in PM (Figure 3). Age‐related trajectories were not modulated by sex differences (p&gt;.05) and were best described by a linear relationship according to model comparison testing different polynomials.ConclusionBy encompassing the entire adult lifespan and assessing longitudinal changes, our study provides strong support that the two MTL networks display differential alterations in the context of aging, as well as sex differences. Since MTL connectivity has been described as particularly affected in the early stages of AD, and age and sex are risk factors, these findings may be helpful to enhance the clinical utility of such biomarkers for the early detection of AD‐related changes.

Differential effects of age and sex on medial temporal lobe networks revealed by longitudinal analyses across the adult lifespan

AbstractBackgroundThe medial temporal lobe (MTL) is extensively connected to the rest of the brain through two specific networks which are particularly affected in Alzheimer’s disease (AD): the anterior‐temporal (AT) and posterior‐medial (PM) systems. As the specific and respective effects of age and sex on the functional integrity of these networks are still largely unknown, we sought to fill this gap using longitudinal data spanning the entire adult life.MethodLongitudinal data from 215 cognitively unimpaired adults (aged 19‐85, including 113 females, followed‐up up to 47 months, 419 scans) were analysed. Seed‐based analyses were applied to resting‐state functional magnetic resonance imaging to generate connectivity maps (seeds: the perirhinal cortex for AT and the parahippocampal cortex for PM). The AT and PM networks were defined by comparing baseline maps from adults younger than 40 using a paired permutation t‐test with threshold‐free cluster enhancement (TFCE) and familywise error (FWE) correction (p&lt;.05) (Figure 1). These masks were then used to extract the mean functional connectivity (FC) for each participant. The interaction effects between network, age and sex were investigated using linear mixed models.ResultThere was a significant interaction between age and networks (p&lt;.001), with a decrease of FC with age in PM and an increase in AT (Figure 2). Sex differences were also observed within these two systems (p&lt;.001), with males exhibiting higher FC than females in AT but lower FC in PM (Figure 3). Age‐related trajectories were not modulated by sex differences (p&gt;.05) and were best described by a linear relationship according to model comparison testing different polynomials.ConclusionBy encompassing the entire adult lifespan and assessing longitudinal changes, our study provides strong support that the two MTL networks display differential alterations in the context of aging, as well as sex differences. Since MTL connectivity has been described as particularly affected in the early stages of AD, and age and sex are risk factors, these findings may be helpful to enhance the clinical utility of such biomarkers for the early detection of AD‐related changes.

Model comparison in German stock returns

PurposeIn order to provide an updated view on the drivers of German stock returns, the authors evaluate the relative performance of nine competing neoclassical asset pricing models in the German stock market between November 1991 and December 2021.Design/methodology/approachThe authors conduct asymptotically valid tests of model comparison when the extent of model mispricing is gauged by the squared Sharpe ratio improvement measure of Barillas et al. (2020).FindingsThe study finds that the Fama and French six-factor model with both traditional and updated value factors emerges as the dominant model.Originality/valueThe authors shed new light on the drivers of German stock returns through an updated and extended period of analysis, wider range of potential models and utilization of valid asymptotic tests of model comparison when models are nonnested (Barillas et al., 2020).

Factor Models, Machine Learning, and Asset Pricing

We survey recent methodological contributions in asset pricing using factor models and machine learning. We organize these results based on their primary objectives: estimating expected returns, factors, risk exposures, risk premia, and the stochastic discount factor as well as model comparison and alpha testing. We also discuss a variety of asymptotic schemes for inference. Our survey is a guide for financial economists interested in harnessing modern tools with rigor, robustness, and power to make new asset pricing discoveries, and it highlights directions for future research and methodological advances.

Effect of surcharge loading rate on the performance of surcharge–vacuum preloading with prefabricated horizontal drains

To understand the impact of the surcharge loading rate (SLR) on the prefabricated horizontal drain (PHD)-vacuum preloading reinforcement dredged fills, four sets of indoor model comparison tests were carried out to analyze and compare the soil surface settlement, lateral displacement, displacement, pore water pressure, vane shear strength, and water content. The degree of consolidation and bearing capacity calculations show that the surcharge loading–PHD-vacuum preloading scheme can effectively improve the soil reinforcement effect, effectively restrain the lateral displacement of the soil during the consolidation process, and improve its stability. The greater the SLR, the better the reinforcement effect and the smaller the lateral displacement; however, this increase is not significant. The optimal SLR found in this study is 4 kPa/d.

Verification Testing of Multi-Dynamical Solver for Multiferroic Antennas

Verification testing of an unconditionally stable finite-difference time-domain (FDTD) solver is presented to validate the accuracy of a numerical framework to understand 3D multiferroic antennas, i.e., a coupled mechanics and electromagnetic problem. Within this framework, the electric, mechanical, and magnetic constitutive relations are substituted into Maxwell's equations and Newton's laws, with the Gilbert damping included within a rate dependent magnetic constitutive relation. Three model comparison tests are performed, one for each domain of physics that the algorithm covers, i.e., electrodynamics, elastodynamics, and micromagnetics damping.

The scaled target learning model: Revisiting learning in the balloon analogue risk task

The Balloon Analogue Risk Task (BART) is a sequential decision making paradigm that assesses risk-taking behavior. Several computational models have been proposed for the BART that characterize risk-taking propensity. An aspect of task performance that has proven challenging to model is the learning that develops from experiencing wins and losses across trials, which has the potential to provide further insight into risky decision making. We developed the Scaled Target Learning (STL) model for this purpose. STL describes learning as adjustments to an individual’s strategy in reaction to outcomes in the task, with the size of adjustments reflecting an individual’s sensitivity to wins and losses. STL is shown to be sensitive to the learning elicited by experimental manipulations. In addition, the model matches or bests the performance of three competing models in traditional model comparison tests (e.g., parameter recovery performance, predictive accuracy, sensitivity to risk-taking propensity). Findings are discussed in the context of the learning process involved in the task. By characterizing the extent to which people are willing to adapt their strategies based on past experience, STL is a step toward a complete depiction of the psychological processes underlying sequential risk-taking behavior.

Comparing Asset Pricing Factor Models under Multivariate t-Distribution: Evidence from China

Factor models provide a cornerstone for understanding financial asset pricing; however, research on China’s stock market risk premia is still limited. Motivated by this, this paper proposes a four-factor model for China’s stock market that includes a market factor, a size factor, a value factor, and a liquidity factor. We compare our four-factor model with a set of prominent factor models based on newly developed likelihood-ratio tests and Bayesian methods. Along with the comparison, we also find supporting evidence for the alternative t-distribution assumption for empirical asset pricing studies. Our results show the following: (1) distributional tests suggest that the returns of factors and stock return anomalies are fat-tailed and therefore are better captured by t-distributions than by normality; (2) under t-distribution assumptions, our four-factor model outperforms a set of prominent factor models in terms of explaining the factors in each other, pricing a comprehensive list of stock return anomalies, and Bayesian marginal likelihoods; (3) model comparison results vary across normality and t-distribution assumptions, which suggests that distributional assumptions matter for asset pricing studies. This paper contributes to the literature by proposing an effective asset pricing factor model and providing factor model comparison tests under non-normal distributional assumptions in the context of China.

Cross-sectional tests of asset pricing models with full-rank mimicking portfolios

In the standard tests of asset pricing models, factor risk premia are estimated on a test asset span so that models are tested with degrees of freedom reduced by the number of factors. Risk premia of traded factors can be further restricted to be equal to their expected returns, but such restrictions cannot be imposed on models with nontraded factors, which may create a problem of testing without full restrictions or on unequal asset spans across models. We propose a full-rank mimicking portfolio approach by projecting nontraded factors onto a combined span of test assets and benchmark traded factors. Under the Hansen-Jagannathan distance framework, we demonstrate that full-rank mimicking portfolios can provide improved power and fair performance comparison against a benchmark model in both specification and model comparison tests.