Reduced-rank clustered coefficient regression for addressing multicollinearity in heterogeneous coefficient estimation.

Clustered coefficient regression (CCR) extends the classical regression model by allowing regression coefficients varying across observations and forming clusters of observations. It has become an increasingly useful tool for modeling the heterogeneous relationship between the predictor and response variables. A typical issue of existing CCR methods is that the estimation and clustering results can be unstable in the presence of multicollinearity. To address the instability issue, this paper introduces a low-rank structure of the CCR coefficient matrix and proposes a penalized non-convex optimization problem with an adaptive group fusion-type penalty tailor-made for this structure. An iterative algorithm is developed to solve this non-convex optimization problem with guaranteed convergence. An upper bound for the coefficient estimation error is also obtained to show the statistical property of the estimator. Empirical studies on both simulated datasets and a COVID-19 mortality rate dataset demonstrate the superiority of the proposed method to existing methods.

Predictive biomarkers for embryotoxicity: a machine learning approach to mitigating multicollinearity in RNA-Seq.

Multicollinearity, characterized by significant co-expression patterns among genes, often occurs in high-throughput expression data, potentially impacting the predictive model's reliability. This study examined multicollinearity among closely related genes, particularly in RNA-Seq data obtained from embryoid bodies (EB) exposed to 5-fluorouracil perturbation to identify genes associated with embryotoxicity. Six genes-Dppa5a, Gdf3, Zfp42, Meis1, Hoxa2, and Hoxb1-emerged as candidates based on domain knowledge and were validated using qPCR in EBs perturbed by 39 test substances. We conducted correlation studies and utilized the variance inflation factor (VIF) to examine the existence of multicollinearity among the genes. Recursive feature elimination with cross-validation (RFECV) ranked Zfp42 and Hoxb1 as the top two among the seven features considered, identifying them as potential early embryotoxicity assessment biomarkers. As a result, a t test assessing the statistical significance of this two-feature prediction model yielded a p value of 0.0044, confirming the successful reduction of redundancies and multicollinearity through RFECV. Our study presents a systematic methodology for using machine learning techniques in transcriptomics data analysis, enhancing the discovery of potential reporter gene candidates for embryotoxicity screening research, and improving the predictive model's predictive accuracy and feasibility while reducing financial and time constraints.

Socio-demographic and infrastructural variables influencing maternal risk concentration among ever-married women of reproductive age in rural West Bengal, India.

BACKGROUND: The risk of a woman dying as a result of pregnancy or childbirth during her lifetime is about one in six in the poorest parts of the world. OBJECTIVES: The present study aims to determine prevalence of maternal risk and the influencing variables among ever-married women belonging to the reproductive age group (15-49) of Birbhum district, West Bengal. METHODS: A cohort-based retrospective cross-sectional study was carried out among the sample of 229 respondents through a purposive stratified random sampling method and a pre-designed semi-structured questionnaire. The ordinal logistic regression (OLR) model was taken as a tool of assessment. Before developing the proportional OLR model, we have checked the multicollinearity effect among the predictors and the first-order effect modifier was evaluated as well. We performed data analysis using SPSS version 26. RESULTS: The result shows that illiterate women (Odds ratios [OR] = 2.81, 95% CI, 0.277-1.791), from lower standard of living (OR = 1.14, 95% CI, -0.845-1.116), married before the age of 15 years (OR = 21.96, 95% CI, -0.55-6.73) and between the age of 15-18 years (OR = 24.51. 95% CI, -0.45-6.85) are more likely to be affected by the higher concentration of maternal risk. Other important predictor is the time of pregnancy registration. Considering the transport and related en-route causalities, the result portraying a clear picture where the distance and travel time becoming significant factors in determining the concentration of maternal risk. CONCLUSION: Incidences of child marriages should be restricted. Eradicating factors influencing an individual's decision to seek care would be an essential contribution in excluding the dominant maternal risk factors.

Understanding the Consequences of Collinearity for Multilevel Models: The Importance of Disaggregation Across Levels.

In multilevel models, disaggregating predictors into level-specific parts (typically accomplished via centering) benefits parameter estimates and their interpretations. However, the importance of level-specificity has been sparsely addressed in multilevel literature concerning collinearity. In this study, we develop novel insights into the interactivity of centering and collinearity in multilevel models. After integrating the broad literatures on centering and collinearity, we review level-specific and conflated correlations in multilevel data. Next, by deriving formal relationships between predictor collinearity and multilevel model estimates, we demonstrate how the consequences of collinearity change across different centering specifications and identify data characteristics that may exacerbate or mitigate those consequences. We show that when all or some level-1 predictors are uncentered, slope estimates can be greatly biased by collinearity. Disaggregation of all predictors eliminates the possibility that fixed effect estimates will be biased due to collinearity alone; however, under some data conditions, collinearity is associated with biased standard errors and random effect (co)variance estimates. Finally, we illustrate the importance of disaggregation for diagnosing collinearity in multilevel data and provide recommendations for the use of level-specific collinearity diagnostics. Overall, the necessity of disaggregation for identifying and managing collinearity's consequences in multilevel models is clarified in novel ways.

An Aeromagnetic Compensation Strategy for Large UAVs.

Aeromagnetic surveys are widely used in geological exploration, mineral resource assessment, environmental monitoring, military reconnaissance, and other areas. It is necessary to perform magnetic compensation for interference in these fields. In recent years, large unmanned aerial vehicles (UAVs) have been more suitable for magnetic detection missions because of the greater loads they can carry. This article proposes some methods for the magnetic compensation of large multiload UAVs. Because of the interference of the large platform and instrument noise, the standard deviations (stds) of the compensation data used in this paper are larger. At the beginning of this article, using the traditional T-L model, we avoid the shortcomings of the anti-magnetic interference ability of triaxial magnetic gate magnetometers. The direction cosine information is obtained by using an inertial navigation system, the global positioning system, and a triaxial magnetic gate magnetometer. Then, we increase the amplitude of the maneuvers in the compensation process; this reduces the multicollinearity problems in the compensation matrix to a certain extent, but it also results in greater magnetic field interference. Lastly, we employ the method of Lasso regularization Newton iteration (LRNM). Compared to the traditional methods of least squares (LS) and singular value decomposition (SVD), LRNM provides improvements of 34% and 27%, respectively. In summary, this series of schemes can be used to perform effective compensation for large multi-load UAVs and improve the actual use of large UAVs, making them more accurate in the measurement of aeromagnetic survey data.

Exploring body morphometry and weight prediction in Ganjam goats in India through principal component analysis.

The body conformations of 262 adult Ganjam goats were subjected to principal component analysis (PCA) with 11 morphometric variables. The results were then used to predict the mature body weight of the goats. Most of the traits were positively correlated, and the correlations were statistically significant. The three main components that the PCA recovered explained 76.12% of the variation in body morphometry overall. The first component accounted for approximately 54.74% of the overall variation and described almost all the traits except ear length and tail length, as indicated by high component loadings. The second component accounted for approximately 11.48% of the variation, mostly accounting for the variation in tail length. The principal component accounted for 9.89% and mostly explained the variation in ear length. The communalities ranged between 0.557 (horn length) and 0.848 (chest circumference) for the first three extracted components. The highest percentage of variability in chest girth was explained by the first three principal components, whereas it was the lowest for the horn length. In the context of predicting body weight through stepwise regression analysis, nine primary variables accounted for 57.3% of the total variance in body weight. Conversely, utilizing the first principal component alongside six additional principal components as independent variables resulted in capturing 56.3% of the variation in the adult live weight of goats while maintaining model comparability with other pertinent parameters. PCA was used efficiently for body weight prediction from major morphometric traits of Ganjam goats addressing the multicollinearity issue.

Machine learning approach for the detection of vitamin D level: a comparative study.

BACKGROUND: After the World Health Organization declared the COVID-19 pandemic, the role of Vitamin D has become even more critical for people worldwide. The most accurate way to define vitamin D level is 25-hydroxy vitamin D(25-OH-D) blood test. However, this blood test is not always feasible. Most data sets used in health science research usually contain highly correlated features, which is referred to as multicollinearity problem. This problem can lead to misleading results and overfitting problems in the ML training process. Therefore, the proposed study aims to determine a clinically acceptable ML model for the detection of the vitamin D status of the North Cyprus adult participants accurately, without the need to determine 25-OH-D level, taking into account the multicollinearity problem. METHOD: The study was conducted with 481 observations who applied voluntarily to Internal Medicine Department at NEU Hospital. The classification performance of four conventional supervised ML models, namely, Ordinal logistic regression(OLR), Elastic-net ordinal regression(ENOR), Support Vector Machine(SVM), and Random Forest (RF) was compared. The comparative analysis is performed regarding the model's sensitivity to the participant's metabolic syndrome(MtS)'positive status, hyper-parameter tuning, sensitivities to the size of training data, and the classification performance of the models. RESULTS: Due to the presence of multicollinearity, the findings showed that the performance of the SVM(RBF) is obviously negatively affected when the test is examined. Moreover, it can be obviously detected that RF is more robust than other models when the variations in the size of training data are examined. This experiment's result showed that the selected RF and ENOR showed better performances than the other two models when the size of training samples was reduced. Since the multicollinearity is more severe in the small samples, it can be concluded that RF and ENOR are not affected by the presence of the multicollinearity problem. The comparative analysis revealed that the RF classifier performed better and was more robust than the other proposed models in terms of accuracy (0.94), specificity (0.96), sensitivity or recall (0.94), precision (0.95), F1-score (0.95), and Cohen's kappa (0.90). CONCLUSION: It is evident that the RF achieved better than the SVM(RBF), ENOR, and OLR. These comparison findings will be applied to develop a Vitamin D level intelligent detection system for being used in routine clinical, biochemical tests, and lifestyle characteristics of individuals to decrease the cost and time of vitamin D level detection.

Novel spatial models for analysis the long-term impact of LULC changes on hydrological components at sub-basin level.

The main objective of this research is to assess the impacts land use and land cover changes (LULC) on hydrological components using novel spatial models at sub-basin scales. The Soil and Water Assessment Tool (SWAT) was employed to analyze the long-term effect of LULC on hydrological components. The results of the calibrated and validated SWAT model demonstrated that run-off and actual evapotranspiration (ET) are expected to experience the largest increase, more than 130% and 90% in autumn, whereas the largest decrease is anticipated to occur in the summer and winter for potential evapotranspiration (PET) (-59%) and ET (-80%) by the projected time. The impacts of hydrological components, elevation, LULC, and an indicator of urbanization and land-use intensity (La) on water yield (WYLD) at sub-basin levels were then considered by four novel spatial models due to the problem of multicollinearity which is prevalent in traditional models. In particular, the Moran eigenvector spatially varying coefficients (MESVC) showed that the soil class out of LULC categories and lateral flow among hydrological properties are expected to have a statistically significant effect on spatial fluctuation of WYLD at the sub-basin scale. The results of spatially filtered unconditional quantile regression (SF-UQR) confirm the findings of the MESVC model and further implied that the lateral flow remains as a statistically significant contributor to WYLD only in lower quantiles (e.g., for quantiles lower than 0.25). The impacts of LULCs on WYLD were statistically lower than the effects caused by the hydrological components.

Predicting mobility using limited data during early stages of a pandemic.

The COVID-19 pandemic has changed consumer behavior substantially. In this study, we explore the drivers of consumer mobility in several metropolitan areas in the United States under the perceived risks of COVID-19. We capture multiple dimensions of perceived risk using local and national cases and death counts of COVID-19, along with real-time Google Trends data for personal protective equipment (PPE). While Google Trends data are popular inputs in many studies, the risk of multicollinearity escalates with the addition of more relevant terms. Therefore, multicollinearity-alleviating methods are needed to appropriately leverage information provided by Google Trends data. We develop and utilize a novel optimization scheme to induce linear models containing strictly significant covariates and minimal multicollinearity. We find that there are a variety of unique factors that drive mobility in different geographic locations, as well as several factors that are common to all locations.

A hierarchical model for jointly assessing ecological and anthropogenic impacts on animal demography.

The management of sustainable harvest of animal populations is of great ecological and conservation importance. Development of formal quantitative tools to estimate and mitigate the impacts of harvest on animal populations has positively impacted conservation efforts. The vast majority of existing harvest models, however, do not simultaneously estimate ecological and harvest impacts on demographic parameters and population trends. Given that the impacts of ecological drivers are often equal to or greater than the effects of harvest, and can covary with harvest, this disconnect has the potential to lead to flawed inference. In this study, we used Bayesian hierarchical models and a 43-year capture-mark-recovery dataset from 404,241 female mallards Anas platyrhynchos released in the North American midcontinent to estimate mallard demographic parameters. Furthermore, we model the dynamics of waterfowl hunters and habitat, and the direct and indirect effects of anthropogenic and ecological processes on mallard demographic parameters. We demonstrate that density dependence, habitat conditions and harvest can simultaneously impact demographic parameters of female mallards, and discuss implications for existing and future harvest management models. Our results demonstrate the importance of controlling for multicollinearity among demographic drivers in harvest management models, and provide evidence for multiple mechanisms that lead to partial compensation of mallard harvest. We provide a novel model structure to assess these relationships that may allow for improved inference and prediction in future iterations of harvest management models across taxa.