A Review on Variable Selection in Regression Analysis

Multilayer perceptron and support vector regression models for feline parturition date prediction

A crucial challenge in feline obstetric care is the accurate prediction of the parturition date during late pregnancy. The classic simple linear regression (SLR) model, which employed the fetal biparietal diameter (BPD) as the single input feature, was frequently applied for such prediction with limited accuracy. Since Multilayer Perceptron (MLP) and Support Vector Regression (SVR) are now two of the most potent scientific regression models, this study, for the first time, introduced such models as the new promising tools for feline parturition date prediction. The following features were candidate inputs for our models: biparietal diameter (BPD), litter size, and maternal weight. We observed and compared the performance results for each model. As the best-performed model, MLP delivered the highest coefficient score (0.972 ± 0.006), lowest mean absolute error score (1.110 ± 0.060), and lowest mean squared error score (1.540 ± 0.141), respectively. For the first time in this study, BPD, litter size, and maternal weight were considered the essential features for the innovative MLP and SVR modeling. With the optimized model parameters and the described analytical platform, further verification of these advanced models in feline obstetric practices is feasible.

Environmental factors influence cross-talk between a heat shock protein and an oxidative stress protein modification in the lizard Gallotia galloti

Better understanding how organisms respond to their abiotic environment, especially at the biochemical level, is critical in predicting population trajectories under climate change. In this study, we measured constitutive stress biomarkers and protein post-translational modifications associated with oxidative stress in Gallotia galloti, an insular lizard species inhabiting highly heterogeneous environments on Tenerife. Tenerife is a small volcanic island in a relatively isolated archipelago off the West coast of Africa. We found that expression of GRP94, a molecular chaperone protein, and levels of protein carbonylation, a marker of cellular stress, change across different environments, depending on solar radiation-related variables and topology. Here, we report in a wild animal population, cross-talk between the baseline levels of the heat shock protein-like GRP94 and oxidative damage (protein carbonylation), which are influenced by a range of available temperatures, quantified through modelled operative temperature. This suggests a dynamic trade-off between cellular homeostasis and oxidative damage in lizards adapted to this thermally and topologically heterogeneous environment.

A neutral comparison of algorithms to minimize L0 penalties for high-dimensional variable selection

Variable selection methods based on L0 penalties have excellent theoretical properties to select sparse models in a high-dimensional setting. There exist modifications of the Bayesian Information Criterion (BIC) which either control the familywise error rate (mBIC) or the false discovery rate (mBIC2) in terms of which regressors are selected to enter a model. However, the minimization of L0 penalties comprises a mixed-integer problem which is known to be NP-hard and therefore becomes computationally challenging with increasing numbers of regressor variables. This is one reason why alternatives like the LASSO have become so popular, which involve convex optimization problems that are easier to solve. The last few years have seen some real progress in developing new algorithms to minimize L0 penalties. The aim of this article is to compare the performance of these algorithms in terms of minimizing L0 -based selection criteria. Simulation studies covering a wide range of scenarios that are inspired by genetic association studies are used to compare the values of selection criteria obtained with different algorithms. In addition, some statistical characteristics of the selected models and the runtime of algorithms are compared. Finally, the performance of the algorithms is illustrated in a real data example concerned with expression quantitative trait loci (eQTL) mapping.

Automatic Variable Selection Algorithms in Prognostic Factor Research in Neck Pain

This study aims to compare the variable selection strategies of different machine learning (ML) and statistical algorithms in the prognosis of neck pain (NP) recovery. A total of 3001 participants with NP were included. Three dichotomous outcomes of an improvement in NP, arm pain (AP), and disability at 3 months follow-up were used. Twenty-five variables (twenty-eight parameters) were included as predictors. There were more parameters than variables, as some categorical variables had >2 levels. Eight modelling techniques were compared: stepwise regression based on unadjusted p values (stepP), on adjusted p values (stepPAdj), on Akaike information criterion (stepAIC), best subset regression (BestSubset) least absolute shrinkage and selection operator [LASSO], Minimax concave penalty (MCP), model-based boosting (mboost), and multivariate adaptive regression splines (MuARS). The algorithm that selected the fewest predictors was stepPAdj (number of predictors, p = 4 to 8). MuARS was the algorithm with the second fewest predictors selected (p = 9 to 14). The predictor selected by all algorithms with the largest coefficient magnitude was "having undergone a neuroreflexotherapy intervention" for NP (β = from 1.987 to 2.296) and AP (β = from 2.639 to 3.554), and "Imaging findings: spinal stenosis" (β = from -1.331 to -1.763) for disability. Stepwise regression based on adjusted p-values resulted in the sparsest models, which enhanced clinical interpretability. MuARS appears to provide the optimal balance between model sparsity whilst retaining high predictive performance across outcomes. Different algorithms produced similar performances but resulted in a different number of variables selected. Rather than relying on any single algorithm, confidence in the variable selection may be increased by using multiple algorithms.

BR-Net: Band reweighted network for quantitative analysis of rapeseed protein spectroscopy

Compared with the complexity of chemical methods, near-infrared spectroscopy (NIRS) is widely used in the detection of protein content because of its advantages of being fast and non-destructive. Aiming to tackle the problem that the raw near-infrared spectroscopy contains many redundant wavelengths, which affects the accuracy of quantitative prediction and requires expertise to process, we propose an end-to-end network: Band Reweighted Network (BR-Net) that automates wavelength reweighted and quantitative prediction of protein content in rapeseed. Unlike extracting part of wavelengths by the traditional wavelength selection methods, BR-Net retains all spectral wavelengths and assigns different weights to the wavelengths to express the correlation with the corresponding concentration, which enables wavelength selection without ignoring the information contained in the less relevant wavelengths. We compare BR-Net with traditional selection methods such as SPA, LARS, CARS, and UVE to verify its efficiency and robustness, finding that the R² of the training set and test set are 0.9797 and 0.9215, the RMSEC and RMSEP are 0.4053 and 0.8501, respectively, and the RPD is 3.5686, which prove BR-Net outperforms all the traditional methods. The network described here is universally applicable to a variety of NIR quantitative analyses.

A systematic review and evaluation of statistical methods for group variable selection

This review condenses the knowledge on variable selection methods implemented in R and appropriate for datasets with grouped features. The focus is on regularized regressions identified through a systematic review of the literature, following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines. A total of 14 methods are discussed, most of which use penalty terms to perform group variable selection. Depending on how the methods account for the group structure, they can be classified into knowledge and data-driven approaches. The first encompass group-level and bi-level selection methods, while two-step approaches and collinearity-tolerant methods constitute the second category. The identified methods are briefly explained and their performance compared in a simulation study. This comparison demonstrated that group-level selection methods, such as the group minimax concave penalty, are superior to other methods in selecting relevant variable groups but are inferior in identifying important individual variables in scenarios where not all variables in the groups are predictive. This can be better achieved by bi-level selection methods such as group bridge. Two-step and collinearity-tolerant approaches such as elastic net and ordered homogeneity pursuit least absolute shrinkage and selection operator are inferior to knowledge-driven methods but provide results without requiring prior knowledge. Possible applications in proteomics are considered, leading to suggestions on which method to use depending on existing prior knowledge and research question.

Holistic Design of Experiments Using an Integrated Process Model

Statistical experimental designs such as factorial, optimal, or definitive screening designs represent the state of the art in biopharmaceutical process characterization. However, such methods alone do not leverage the fact that processes operate as a mutual interplay of multiple steps. Instead, they aim to investigate only one process step at a time. Here, we want to develop a new experimental design method that seeks to gain information about final product quality, placing the right type of run at the right unit operation. This is done by minimizing the simulated out-of-specification rate of an integrated process model comprised of a chain of regression models that map process parameters to critical quality attributes for each unit operation. Unit operation models are connected by passing their response to the next unit operation model as a load parameter, as is done in real-world manufacturing processes. The proposed holistic DoE (hDoE) method is benchmarked against standard process characterization approaches in a set of in silico simulation studies where data are generated by different ground truth processes to illustrate the validity over a range of scenarios. Results show that the hDoE approach leads to a >50% decrease in experiments, even for simple cases, and, at the same time, achieves the main goal of process development, validation, and manufacturing to consistently deliver product quality.

Using Voice Biomarkers to Classify Suicide Risk in Adult Telehealth Callers: Retrospective Observational Study

BACKGROUND

Artificial intelligence has the potential to innovate current practices used to detect the imminent risk of suicide and to address shortcomings in traditional assessment methods.

OBJECTIVE

In this paper, we sought to automatically classify short segments (40 milliseconds) of speech according to low versus imminent risk of suicide in a large number (n=281) of telephone calls made to 2 telehealth counselling services in Australia.

METHODS

A total of 281 help line telephone call recordings sourced from On The Line, Australia (n=266, 94.7%) and 000 Emergency services, Canberra (n=15, 5.3%) were included in this study. Imminent risk of suicide was coded for when callers affirmed intent, plan, and the availability of means; level of risk was assessed by the responding counsellor and reassessed by a team of clinical researchers using the Columbia Suicide Severity Rating Scale (=5/6). Low risk of suicide was coded for in an absence of intent, plan, and means and via Columbia suicide Severity Scale Ratings (=1/2). Preprocessing involved normalization and pre-emphasis of voice signals, while voice biometrics were extracted using the statistical language r. Candidate predictors were identified using Lasso regression. Each voice biomarker was assessed as a predictor of suicide risk using a generalized additive mixed effects model with splines to account for nonlinearity. Finally, a component-wise gradient boosting model was used to classify each call recording based on precoded suicide risk ratings.

RESULTS

A total of 77 imminent-risk calls were compared with 204 low-risk calls. Moreover, 36 voice biomarkers were extracted from each speech frame. Caller sex was a significant moderating factor (β=-.84, 95% CI -0.85, -0.84; t=6.59, P<.001). Candidate biomarkers were reduced to 11 primary markers, with distinct models developed for men and women. Using leave-one-out cross-validation, ensuring that the speech frames of no single caller featured in both training and test data sets simultaneously, an area under the precision or recall curve of 0.985 was achieved (95% CI 0.97, 1.0). The gamboost classification model correctly classified 469,332/470,032 (99.85%) speech frames.

CONCLUSIONS

This study demonstrates an objective, efficient, and economical assessment of imminent suicide risk in an ecologically valid setting with potential applications to real-time assessment and response.

TRIAL REGISTRATION

Australian New Zealand Clinical Trials Registry ACTRN12622000486729; https://www.anzctr.org.au/ACTRN12622000486729.aspx.

Extrapolation Assessment for Forest Structural Parameters in Planted Forests of Southern China by UAV-LiDAR Samples and Multispectral Satellite Imagery

Accurate estimation and extrapolation of forest structural parameters in planted forests are essential for monitoring forest resources, investigating their ecosystem services (e.g., forest structure and functions), as well as supporting decisions for precision silviculture. Advances in unmanned aerial vehicle (UAV)-borne Light Detection and Ranging (LiDAR) technology have enhanced our ability to precisely characterize the 3-D structure of the forest canopy with high flexibility, usually within forest plots and stands. For wall-to-wall forest structure mapping in broader landscapes, samples (transects) of UAV-LiDAR datasets are a cost-efficient solution as an intermediate layer for extrapolation from field plots to full-coverage multispectral satellite imageries. In this study, an advanced two-stage extrapolation approach was established to estimate and map large area forest structural parameters (i.e., mean DBH, dominant height, volume, and stem density), in synergy with field plots and UAV-LiDAR and GF-6 satellite imagery, in a typical planted forest of southern China. First, estimation models were built and used to extrapolate field plots to UAV-LiDAR transects; then, the maps of UAV-LiDAR transects were extrapolated to the whole study area using the wall-to-wall grid indices that were calculated from GF-6 satellite imagery. By comparing with direct prediction models that were fitted by field plots and GF-6-derived spectral indices, the results indicated that the two-stage extrapolation models (R2 = 0.64–0.85, rRMSE = 7.49–26.85%) obtained higher accuracy than direct prediction models (R2 = 0.58–0.75, rRMSE = 21.31–38.43%). In addition, the effect of UAV-LiDAR point density and sampling intensity for estimation accuracy was studied by sensitivity analysis as well. The results showed a stable level of accuracy for approximately 10% of point density (34 pts·m−2) and 20% of sampling intensity. To understand the error propagation through the extrapolation procedure, a modified U-statistics uncertainty analysis was proposed to characterize pixel-level estimates of uncertainty and the results demonstrated that the uncertainty was 0.75 cm for mean DBH, 1.23 m for dominant height, 14.77 m3·ha−1 for volume and 102.72 n·ha−1 for stem density, respectively.

Study on the Water and Heat Fluxes of a Very Humid Forest Ecosystem and Their Relationship with Environmental Factors in Jinyun Mountain, Chongqing

The high-humidity mountain forest ecosystem (HHMF) of Jinyun Mountain in Chongqing is a fragile ecosystem that is sensitive to climate change and human activities. Because it is shrouded in fog year-round, illumination in the area is seriously insufficient. However, the flux (energy, water) exchanges (FEs) in this ecosystem and their influencing factors are not clear. Using one-year data from flux towers with a double-layer (25 m and 35 m) eddy covariance (EC) observation system, we proved the applicability of the EC method on rough underlying surfaces, quantified the FEs of HHMFs, and found that part of the fog might also be observed by the EC method. The observation time was separated from day and night, and then the environmental control of the FEs was determined by stepwise regression analysis. Through the water balance, it was proven that the negative value of evapotranspiration (ETN), which represented the water vapor input from the atmosphere to the ecosystem, could not be ignored and provided a new idea for the possible causes of the evaporation paradox. The results showed that the annual average daily sensible heat flux (H) and latent heat flux (LE) ranged from −126.56 to 131.27 W m−2 and from −106.7 to 222.27 W m−2, respectively. The annual evapotranspiration (ET), positive evapotranspiration (ETP), and negative evapotranspiration (ETN) values were 389.31, 1387.76, and −998.45 mm, respectively. The energy closure rate of the EC method in the ecosystems was 84%. Fog was the ETN observed by the EC method and an important water source of the HHMF. Therefore, the study area was divided into subtropical mountain cloud forests (STMCFs). Stepwise regression analysis showed that the H and LE during the day were mainly determined by radiation (Rn) and temperature (Tair), indicating that the energy of the ecosystem was limited, and future climate warming may enhance the FEs of the ecosystem. Additionally, ETN was controlled by wind speed (WS) in the whole period, and WS was mainly affected by altitude and temperature differences within the city. Therefore, fog is more likely to occur in the mountains near heat island cities in tropical and subtropical regions. This study emphasizes that fog, as an important water source, is easily ignored in most EC methods and that there will be a large amount of fog in ecosystems affected by future climate warming, which can explain the evaporation paradox.

Accurate Prediction of Persistent Upper Extremity Impairment in Patients With Ischemic Stroke

OBJECTIVE

To develop a simple and effective risk score for predicting which stroke patients will have persistent impairment of upper extremity motor function at 90 days.

DESIGN

Post hoc analysis of clinical trial patients hospitalized with acute ischemic stroke who were followed for 90 days to determine functional outcome.

SETTING

Patient were hospitalized at facilities across the United States.

PARTICIPANTS

We created a harmonized cohort of individual patients (N=1653) from the NINDS tPA, ALIAS part 2, IMS-III, DEFUSE 3, and FAST-MAG trials. We split the cohort into balanced derivation and validation samples.

INTERVENTIONS

Not applicable.

MAIN OUTCOME MEASURES

The primary outcome was persistent arm impairment, defined as a National Institutes of Health Stroke Scale (NIHSS) arm domain score of 2 to 4 at 90 days in patients who had a 24-hour NIHSS arm score of 1 or more. We used least absolute shrinkage and selection operator regression to determine the elements of the persistent upper extremity impairment (PUPPI) index, which we validated as a predictive tool.

RESULTS

We included 1653 patients (827 derivation, 826 validation), of whom 803 (48.6%) had persistent arm impairment. The PUPPI index gives 1 point each for age 55 years or older and NIHSS values of worse arm (4), worse leg (>2), facial palsy (3), and total NIHSS (≥10). The optimal cutpoint for the PUPPI index was 3 or greater, at which the area under the curve was greater than 0.75 for the derivation and validation cohorts and when using NIHSS values from either 24 hours or in a subacute or discharge time window. Results were similar across different levels of stroke severity.

CONCLUSION

The PUPPI index uses readily available information to accurately predict persistent upper extremity motor impairment at 90 days poststroke. The PUPPI index can be administered in minutes and could be used as inclusion criterion in recovery-related clinical trials or, with additional development, as a prognostic tool for patients, caregivers, and clinicians.

Comparison of standard and penalized logistic regression in risk model development

Objective

Regression models are ubiquitous in thoracic surgical research. We aimed to compare the value of standard logistic regression with the more complex but increasingly used penalized regression models using a recently published risk model as an example.

Methods

Using a standardized data set of clinical T1-3N0 esophageal cancer patients, we created models to predict the likelihood of unexpected pathologic nodal disease after surgical resection. Models were fitted using standard logistic regression or penalized regression (ridge, lasso, elastic net, and adaptive lasso). We compared the model performance (Brier score, calibration slope, C statistic, and overfitting) of standard regression with penalized regression models.

Results

Among 3206 patients with clinical T1-3N0 esophageal cancer, 668 (22%) had unexpected pathologic nodal disease. Of the 15 candidate variables considered in the models, the key predictors of nodal disease included clinical tumor stage, tumor size, grade, and presence of lymphovascular invasion. The standard regression model and all 4 penalized logistic regression models had virtually identical performance with Brier score ranging from 0.138 to 0.141, concordance index ranging from 0.775 to 0.788, and calibration slope from 0.965 to 1.05.

Conclusions

For predictive modeling in surgical outcomes research, when the data set is large and the outcome of interest is relatively frequent, standard regression models and the more complicated penalized models are very likely to have similar predictive performance. The choice of statistical methods for risk model development should be on the basis of the nature of the data at hand and good statistical practice, rather than the novelty or complexity of statistical models.

Saudi Non-Oil Exports before and after COVID-19: Historical Impacts of Determinants and Scenario Analysis

The diversification of the economy including its exports is at the core of Saudi Vision 2030. The vision targets to raise non-oil export from 16% to 50% of non-oil GDP by 2030. Achieving this, in addition to other goals, necessitates a better understanding of the non-oil export relationship with its determinants. However, we are not aware of a study that estimates the impacts of the determinants on Saudi non-oil exports covering the recent years of reforms and low oil prices and that conducts simulations for future. The purpose of this study is to develop an econometric modeling framework for Saudi non-oil export that can enhance informing the policymaking process through empirical estimations and simulations. For estimations, we applied cointegration and equilibrium correction methodology to the annual data for the period 1983–2018. Results show that Middle Eastern and North African countries’ GDP, as a measure of foreign income, and Saudi Arabia’s non-oil GDP, as a measure of production capacity, have statistically significant positive effects on Saudi non-oil exports in the long run. The real effective exchange rate (REER), as a measure of competitiveness, also exerts a positive effect in the long run if it depreciates and vice versa. Furthermore, our findings support the Export-led growth concept, which articulates that export can be an engine of economic growth and does not support the Dutch disease concept, which highlights the consequences of the resource sector for the non-resource tradable sector for Saudi Arabia. Macroeconometric model-based simulations conducted up to 2030 reveal out that the Saudi non-oil export is more responsive to the changes in REER than any other determinants. The simulation results also show that non-oil manufacturing makes a three times larger contribution to the future expansion of non-oil exports than agriculture. Moreover, the simulations discover that finance, insurance, and other business services, as well as transport and communication play an important role in improving the Saudi non-oil export performance in the coming decade. The key policy recommendation is that measures should be implemented in a coordinated and balanced way to achieve non-oil exports and other targets of the Vision.

Association of State Social and Environmental Factors With Rates of Self-injury Mortality and Suicide in the United States

IMPORTANCE

Self-injury mortality (SIM) combines suicides and the preponderance of drug misuse-related overdose fatalities. Identifying social and environmental factors associated with SIM and suicide may inform etiologic understanding and intervention design.

OBJECTIVE

To identify factors associated with interstate SIM and suicide rate variation and to assess potential for differential suicide misclassification.

DESIGN, SETTING, AND PARTICIPANTS

This cross-sectional study used a partial panel time series with underlying cause-of-death data from 50 US states and the District of Columbia for 1999-2000, 2007-2008, 2013-2014 and 2018-2019. Applying data from the Centers for Disease Control and Prevention, SIM includes all suicides and the preponderance of unintentional and undetermined drug intoxication deaths, reflecting self-harm behaviors. Data were analyzed from February to June 2021.

EXPOSURES

Exposures included inequity, isolation, demographic characteristics, injury mechanism, health care access, and medicolegal death investigation system type.

MAIN OUTCOMES AND MEASURES

The main outcome, SIM, was assessed using unstandardized regression coefficients of interstate variation associations, identified by the least absolute shrinkage and selection operator; ratios of crude SIM to suicide rates per 100 000 population were assessed for potential differential suicide misclassification.

RESULTS

A total of 101 325 SIMs were identified, including 74 506 (73.5%) among males and 26 819 (26.5%) among females. SIM to suicide rate ratios trended upwards, with an accelerating increase in overdose fatalities classified as unintentional or undetermined (SIM to suicide rate ratio, 1999-2000: 1.39; 95% CI, 1.38-1.41; 2018-2019: 2.12; 95% CI, 2.11-2.14). Eight states recorded a SIM to suicide rate ratio less than 1.50 in 2018-2019 vs 39 states in 1999-2000. Northeastern states concentrated in the highest category (range, 2.10-6.00); only the West remained unrepresented. Least absolute shrinkage and selection operator identified 8 factors associated with the SIM rate in 2018-2019: centralized medical examiner system (β = 4.362), labor underutilization rate (β = 0.728), manufacturing employment (β = -0.056), homelessness rate (β = -0.125), percentage nonreligious (β = 0.041), non-Hispanic White race and ethnicity (β = 0.087), prescribed opioids for 30 days or more (β = 0.117), and percentage without health insurance (β = -0.013) and 5 factors associated with the suicide rate: percentage male (β = 1.046), military veteran (β = 0.747), rural (β = 0.031), firearm ownership (β = 0.030), and pain reliever misuse (β = 1.131).

CONCLUSIONS AND RELEVANCE

These findings suggest that SIM rates were associated with modifiable, upstream factors. Although embedded in SIM, suicide unexpectedly deviated in proposed social and environmental determinants. Heterogeneity in medicolegal death investigation processes and data assurance needs further characterization, with the goal of providing the highest-quality reports for developing and tracking public health policies and practices.

Application of Metaheuristic Approaches for the Variable Selection Problem

Variable selection is an old topic from regression models. Besides many conventional approaches, some metaheuristic approaches from the realm of optimization such as GA (Genetic Algorithm) or simulated annealing have been suggested to date. These methods have a considerable advantage to deal with many problems over the classical methods, but they must control relevant fine-tuning parameters associated with cross-over or mutation, which can be difficult and time-consuming. In this paper, Jaya, one of several parameter-free approaches will be suggested and explored. Several metaheuristic methods will be compared using results from a real-world dataset and a simulated dataset. The impact of using local search will be analyzed.

Prestroke Disability and Outcome After Thrombectomy for Emergent Anterior Circulation Large Vessel Occlusion Stroke

BACKGROUND AND OBJECTIVES

To determine the impact of endovascular therapy for large vessel occlusion stroke in patients with vs those without premorbid disability.

METHODS

We performed a post hoc analysis of the TREVO Stent-Retriever Acute Stroke (TRACK) Registry, which collected data on 634 consecutive patients with stroke treated with the Trevo device as first-line endovascular thrombectomy (EVT) at 23 centers in the United States. We included patients with internal carotid or middle cerebral (M1/M2 segment) artery occlusions, and the study exposure was patient- or caregiver-reported premorbid modified Rank Scale (mRS) score ≥2 (premorbid disability [PD]) vs premorbid mRS score of 0 to 1 (no PD [NPD]). The primary outcome was no accumulated disability, defined as no increase in 90-day mRS score from the patient's premorbid mRS score.

RESULTS

Of the 634 patients in TRACK, 407 patients were included in our cohort, of whom 53 (13.0%) had PD. The primary outcome of no accumulated disability was achieved in 37.7% (20 of 53) of patients with PD and 16.7% (59 of 354) of patients with NPD (p < 0.001), while death occurred in 39.6% (21 of 53) and 14.1% (50 of 354) (p < 0.001), respectively. The adjusted odds ratio of no accumulated disability for patients with PD was 5.2 (95% confidence interval [CI] 2.4-11.4, p < 0.001) compared to patients with NPD. However, the adjusted odds ratio for death in patients with PD was 2.90 (95% CI 1.38-6.09, p = 0.005).

DISCUSSION

In this study of patients with anterior circulation acute ischemic stroke treated with EVT, we found that PD was associated with a higher probability of not accumulating further disability compared to patients with NPD but also with higher probability of death.

CLASSIFICATION OF EVIDENCE

This study provides Class II evidence that in anterior circulation acute ischemic stroke treated with EVT, patients with PD compared to those without disability were more likely not to accumulate more disability but were more likely to die.

Identifying stress responsive genes using overlapping communities in co-expression networks

Background

This paper proposes a workflow to identify genes that respond to specific treatments in plants. The workflow takes as input the RNA sequencing read counts and phenotypical data of different genotypes, measured under control and treatment conditions. It outputs a reduced group of genes marked as relevant for treatment response. Technically, the proposed approach is both a generalization and an extension of WGCNA. It aims to identify specific modules of overlapping communities underlying the co-expression network of genes. Module detection is achieved by using Hierarchical Link Clustering. The overlapping nature of the systems’ regulatory domains that generate co-expression can be identified by such modules. LASSO regression is employed to analyze phenotypic responses of modules to treatment.

Results

The workflow is applied to rice (Oryza sativa), a major food source known to be highly sensitive to salt stress. The workflow identifies 19 rice genes that seem relevant in the response to salt stress. They are distributed across 6 modules: 3 modules, each grouping together 3 genes, are associated to shoot K content; 2 modules of 3 genes are associated to shoot biomass; and 1 module of 4 genes is associated to root biomass. These genes represent target genes for the improvement of salinity tolerance in rice.

Conclusions

A more effective framework to reduce the search-space for target genes that respond to a specific treatment is introduced. It facilitates experimental validation by restraining efforts to a smaller subset of genes of high potential relevance.

The effect of web of science subject categories on clustering: the case of data-driven methods in business and economic sciences

The primary goal of this article is to identify the research fronts on the application of data-driven methods in business and economics. For this purpose, the research literature of the business and economic sciences Subject Categories from the Web of Science is mapped using BibExcel and VOSviewer. Since the assignment to subject categories is done at the journal level and since a journal is often assigned to several subject categories in Web of Science, two mappings are performed: one without considering multiple assignments (broad view) and one considering only those (articles from) journals that have been assigned exclusively to the business and economic sciences subject categories and no others (narrow view). A further aim of this article is therefore to identify differences in the two mappings. Surprisingly, engineering sciences play a major role in the broad mapping, in addition to the economic sciences. In the narrow mapping, however, only the following clusters with a clear business-management focus emerge: (i) Data-driven methods in management in general and data-driven supply chain management in particular, (ii) Data-driven operations research analyses with different business administration/management focuses, (iii) Data-driven methods and processes in economics and finance, and (iv) Data-driven methods in Information Systems. One limitation of the narrow mapping is that many relevant documents are not covered since the journals in which they appear are assigned to multiple subject categories in WoS. The paper comes to the conclusion that the multiple assignments of subject categories in Web of Science may lead to massive changes in the results. Adjacent subject areas—in this specific case the application of data-driven methods in engineering and more mathematically oriented contributions in economics (econometrics) are considered in the broad mapping (not excluding subject categories from neighbouring disciplines) and are even over-represented compared to the core areas of business and economics. If a mapping should only consider the core aspects of particular research fields, it is shown in this use case that the exclusion of Web of Science-subject categories that do not belong to the core areas due to multiple assignments (narrow view), may be a valuable alternative. Finally, it depends on the reader to decide which mapping is more beneficial to them.

Variable selection methods were poorly reported but rarely misused in major medical journals: Literature review

Objective This work presents a review of the literature on reporting, practice and misuse of knowledge-based and data-driven variable selection methods, in five highly cited medical journals, considering recoding and interaction unlike previous reviews. Study Design and Setting Original observational studies with a predictive or explicative research question with multivariable analyses published in N. Engl. J. Med., Lancet, JAMA, Br. Med. J. and Ann. Intern. Med. between 2017 and 2019 were searched. Article screening was performed by a single reader, data extraction was performed by two readers and a third reader participated in case of disagreement. The use of data-driven variable selection methods in causal explicative questions was considered as misuse. Results 488 articles were included. The variable selection method was unclear in 234 (48%) articles, data-driven in 78 (16%) articles and knowledge-based in 176 (36%) articles. The most common data-driven methods were: Univariate selection (n = 22, 4.5%) and model comparisons or testing for interaction (n = 17, 3.5%). Data-driven methods were misused in 51 (10.5%) of articles. Conclusion Overall reporting of variable selection methods is insufficient. Data-driven methods seem to be used only in a minority of articles of the big five medical journals.

Association of plasma Aβ40/Aβ42 ratio and brain Aβ accumulation: testing a whole-brain PLS-VIP approach in individuals at risk of Alzheimer's disease

Molecular and brain regional/network-wise pathophysiological changes at preclinical stages of Alzheimer's disease (AD) have primarily been found through knowledge-based studies conducted in late-stage mild cognitive impairment/dementia populations. However, such an approach may compromise the objective of identifying the earliest spatial-temporal pathophysiological processes. We investigated 261 individuals with subjective memory complaints, a condition at increased risk of AD, to test a whole-brain, non-a-priori method based on partial least squares in unraveling the association between plasma Aβ42/Aβ40 ratio and an extensive set of brain regions characterized through molecular imaging of Aβ accumulation and cortical metabolism. Significant associations were mapped onto large-scale networks, identified through an atlas and by knowledge, to elaborate on the reliability of the results. Plasma Aβ42/40 ratio was associated with Aβ-PET uptake (but not FDG-PET) in regions generally investigated in preclinical AD such as those belonging to the default mode network, but also in regions/networks normally not accounted - including the central executive and salience networks - which likely have a selective vulnerability to incipient Aβ accumulation. The present whole-brain approach is promising to investigate early pathophysiological changes of AD to fully capture the complexity of the disease, which is essential to develop timely screening, detection, diagnostic, and therapeutic interventions.

Predicting 5- and 10-Year Mortality Risk in Older Adults With Diabetes

OBJECTIVE

Several diabetes clinical practice guidelines suggest that treatment goals may be modified in older adults on the basis of comorbidities, complications, and life expectancy. The long-term benefits of treatment intensification may not outweigh short-term risks for patients with limited life expectancy. Because of the uncertainty of determining life expectancy for individual patients, we sought to develop and validate prognostic indices for mortality in older adults with diabetes.

RESEARCH DESIGN AND METHODS

We used a prevalence sample of veterans with diabetes who were aged ≥65 years on 1 January 2006 (N = 275,190). Administrative data were queried for potential predictors that included patient demographics, comorbidities, procedure codes, laboratory values and anthropomorphic measurements, medication history, and previous health service utilization. Logistic least absolute shrinkage and selection operator regressions were used to identify variables independently associated with mortality. The resulting odds ratios were then weighted to create prognostic indices of mortality over 5 and 10 years.

RESULTS

Thirty-seven predictors of mortality were identified: 4 demographic variables, prescriptions for insulin or sulfonylureas or blood pressure medications, 6 biomarkers, previous outpatient and inpatient utilization, and 22 comorbidities/procedures. The prognostic indices showed good discrimination, with C-statistics of 0.74 and 0.76 for 5- and 10-year mortality, respectively. The indices also demonstrated excellent agreement between observed outcome and predictions, with calibration slopes of 1.01 for both 5- and 10-year mortality.

CONCLUSIONS

Prognostic indices obtained from administrative data can predict 5- and 10-year mortality in older adults with diabetes. Such a tool may enable clinicians and patients to develop individualized treatment goals that balance risks and benefits of treatment intensification.

International Energy Security Risk Index—Analysis of the Methodological Settings

The main objective of this paper is to analyze model settings of the International Energy Security Risk Index developed by the U.S. Chamber of Commerce. The study was performed using stepwise regression, principal component analysis, and Promax oblique rotation. The conclusion of the regression analysis shows that Crude Oil Price and Global Coal Reserves are sufficient to explain 90% of the variance of the Index. However, if a model that explains 100% of the variance of the Index is chosen and other variables are added, Global Coal Reserves loses importance due to the presence of other parameters in which it is contained. Regardless of the chosen model of analysis, it is evident that there is room for revising the Index and removing variables that do not contribute to its precision. The research showed that the main disadvantage of the variables that make up the Index rests with the fact that the variables are of different degrees of generality, that is, one parameter is contained in other parameters (unclear which other). The research covers data for 25 countries over a 26-year period, with the first year of the research being 1980 and the last 2016 (the latest available report).

BACE and BMA Variable Selection and Forecasting for UK Money Demand and Inflation with Gretl

In this paper, we apply Bayesian averaging of classical estimates (BACE) and Bayesian model averaging (BMA) as an automatic modeling procedures for two well-known macroeconometric models: UK demand for narrow money and long-term inflation. Empirical results verify the correctness of BACE and BMA selection and exhibit similar or better forecasting performance compared with a non-pooling approach. As a benchmark, we use Autometrics—an algorithm for automatic model selection. Our study is implemented in the easy-to-use gretl packages, which support parallel processing, automates numerical calculations, and allows for efficient computations.

Variable selection for inferential models with relatively high-dimensional data: Between method heterogeneity and covariate stability as adjuncts to robust selection

Variable selection in inferential modelling is problematic when the number of variables is large relative to the number of data points, especially when multicollinearity is present. A variety of techniques have been described to identify 'important' subsets of variables from within a large parameter space but these may produce different results which creates difficulties with inference and reproducibility. Our aim was evaluate the extent to which variable selection would change depending on statistical approach and whether triangulation across methods could enhance data interpretation. A real dataset containing 408 subjects, 337 explanatory variables and a normally distributed outcome was used. We show that with model hyperparameters optimised to minimise cross validation error, ten methods of automated variable selection produced markedly different results; different variables were selected and model sparsity varied greatly. Comparison between multiple methods provided valuable additional insights. Two variables that were consistently selected and stable across all methods accounted for the majority of the explainable variability; these were the most plausible important candidate variables. Further variables of importance were identified from evaluating selection stability across all methods. In conclusion, triangulation of results across methods, including use of covariate stability, can greatly enhance data interpretation and confidence in variable selection.