Multivariate Modeling Methods Research Articles

A Multivariate Temperature Drift Modeling and Compensation Method for Large-Diameter High-Precision Fiber Optic Gyroscopes

A Multivariate Temperature Drift Modeling and Compensation Method for Large-Diameter High-Precision Fiber Optic Gyroscopes

Brain-wide functional connectivity patterns support general cognitive ability and mediate effects of socioeconomic status in youth

General cognitive ability (GCA) is an individual difference dimension linked to important academic, occupational, and health-related outcomes and its development is strongly linked to differences in socioeconomic status (SES). Complex abilities of the human brain are realized through interconnections among distributed brain regions, but brain-wide connectivity patterns associated with GCA in youth, and the influence of SES on these connectivity patterns, are poorly understood. The present study examined functional connectomes from 5937 9- and 10-year-olds in the Adolescent Brain Cognitive Development (ABCD) multi-site study. Using multivariate predictive modeling methods, we identified whole-brain functional connectivity patterns linked to GCA. In leave-one-site-out cross-validation, we found these connectivity patterns exhibited strong and statistically reliable generalization at 19 out of 19 held-out sites accounting for 18.0% of the variance in GCA scores (cross-validated partial η2). GCA-related connections were remarkably dispersed across brain networks: across 120 sets of connections linking pairs of large-scale networks, significantly elevated GCA-related connectivity was found in 110 of them, and differences in levels of GCA-related connectivity across brain networks were notably modest. Consistent with prior work, socioeconomic status was a strong predictor of GCA in this sample, and we found that distributed GCA-related brain connectivity patterns significantly statistically mediated this relationship (mean proportion mediated: 15.6%, p < 2 × 10−16). These results demonstrate that socioeconomic status and GCA are related to broad and diffuse differences in functional connectivity architecture during early adolescence, potentially suggesting a mechanism through which socioeconomic status influences cognitive development.

Material manufacturability classification in high shear wet granulation process of Chinese medicine

The high shear wet granulation(HSWG) process of Chinese medicine has a complicated mechanism. There are many influencing factors that contribute to this process. In order to summarize the manufacturability of different kinds of materials in HSWG, this paper constructed a material library composed of 11 materials, including 4 Chinese medicine extracts and 7 pharmaceutical excipients. Each material was described by 22 physical parameters. Several binders were employed, and their density, viscosity and surface tension were characterized. Combining empirical constraints and the principle of randomization, 21 designed experiments and 8 verification experiments were arranged. The partial least squares(PLS) algorithm was used to establish a process model in prediction of the median granule size based on properties of raw materials and binders, and process parameters. The surface tension and density of binders, as well as the maximum pore saturation were identified as key variables. In the latent variable space of the HSWG process model, all materials could be divided into three categories, namely the Chinese medicine extracts, the diluents and the disintegrants. The granulation of Chinese medicine extracts required low viscosity and low amount of binder, and the resulted granule sizes were small. The diluent powders occupied a large physical space, and could be made into granules with different granule sizes by adjusting the properties of binders. The disintegrants tended to be made into large granules under the condition of aqueous binder. The combination use of material database and multivariate modeling method is conducive to innovate the knowledge discovery of the wet granulation process of Chinese medicine, and provides a basis for the formulation and process design based on material attributes.

A multivariate modeling method for the prediction of low fetal fraction before noninvasive prenatal testing.

To investigate factors associated with fetal fraction and to develop a new predictive method for low fetal fraction before noninvasive prenatal testing. The study was a retrospective cohort analysis based on the results of noninvasive prenatal testing, complete blood count, thyroxin test, and Down's syndrome screening during the first or second trimester in 14,043 pregnant women. Random forests algorithm was applied to predict the low fetal fraction status (fetal fraction < 4%) through individual information and laboratory records. The performance of the model was evaluated and compared to predictions using maternal weight. Of 14,043 cases, maternal weight, red blood cell, hemoglobin, and free T3 were significantly negatively correlated with fetal fraction while gestation age, free T4, pregnancy-associated plasma protein-A, alpha-fetoprotein, unconjugated estriol, and β-human chorionic gonadotropin were significantly positively correlated with fetal fraction. Compared to predictions using maternal weight as an isolated parameter, the model had a higher area under the curve of receiver operating characteristic and overall accuracy. The comprehensive predictive method based on combined multiple factors was more effective than a single-factor model in low fetal fraction status prediction. This method can provide more pretest quality control for noninvasive prenatal testing.

Machiavellianism dimensions, religiosity, social environment, and tax evasion

This research aims to investigate the effects of Machiavellianism dimensions that consist of the use of manipulative tactics, cynical views, and disregard of conventional morality on tax evasion. Further, we also analyze the effects of religiosity and social environment as the variables that control Machiavellianism behavior. The population in this research are individual taxpayers, and the sampling technique used is an online survey with the snowball sampling method. The criteria used in sampling are taxpayers who are self-employed or entrepreneurs. The data is then analyzed using the univariate and multivariate General Linear Model (GLM) methods. The results show that manipulative tactics and disregard of conventional morality have positive effects on tax evasion. Meanwhile, cynical views have a negative effect on tax evasion. Further, religiosity has negative effects on the use of manipulative tactics and disregard of conventional morality. Lastly, social environment has a positive effect on cynical views but no significant effects on the use of manipulative tactics and disregard of conventional morality.

INFLUENCE OF SECONDARY CHANGES ON THE COLLECTOR RELIABILITY COEFFICIENT IN GEOLOGICAL MODELS OF OIL AND GAS DEPOSITS

Secondary changes in reservoir rocks, depending on their type, have a significant effect on filtration-capacitive properties (FCP). The result of this effect is a significant complication of the pore space of the reservoir and, as a consequence, the connectivity of the reservoir. The connectivity parameter is an important criterion when choosing multivariate modeling methods and sandstone cube construction parameters when creating geological models. The paper discusses the methods used today to calculate reservoir connectivity in geological models of oil and gas deposits, evaluate their effectiveness and propose a new approach to determining the connectivity parameter depending on the type of secondary changes in reservoir rocks and the geometry of the natural reservoir. In addition, an example shows the dependence of the coupling coefficient on the main parameters of the geological model: sandiness, ranks of variograms and the nugget effect. The proposed method allows to estimate the anisotropy of the main types of secondary changes in various directions and to assess FCP changes in geological models.

Reliability prediction of rotary encoder based on multivariate accelerated degradation modeling

The rotary encoder is an important sensor for measuring rotational motion. A step–stress accelerated degradation test (SSADT) is conducted with respect to accuracy to evaluate the reliability rapidly. The existing errors in output signals are identified as degradation indicator to characterize performance degradation. A methodology for the encoder SSADT modeling is presented. Wiener process theory with unit–to–unit variation is utilized in the model to describe the degradation process of each error. Considering the correlation among the degradation processes, a multivariate degradation modeling method is developed based on the Copula function to estimate the reliability of rotary encoder with multiple degradation indicators. The reliability of the encoder under the use–stress level is derived by combining the margin distributions and Copula function.

Predicting Adaptive Genetic Variation of Loblolly Pine (Pinus taeda L.) Populations Under Projected Future Climates Based on Multivariate Models.

Greenhouse gas emission and global warming are likely to cause rapid climate change within the natural range of loblolly pine over the next few decades, thus bringing uncertainty to their adaptation to the environment. Here, we studied adaptive genetic variation of loblolly pine and correlated genetic variation with bioclimatic variables using multivariate modeling methods-Redundancy Analysis, Generalized Dissimilarity Modeling, and Gradient Forests. Studied trees (N = 299) were originally sampled from their native range across eight states on the east side of the Mississippi River. Genetic variation was calculated using a total of 44,317 single-nucleotide polymorphisms acquired by exome target sequencing. The fitted models were used to predict the adaptive genetic variation on a large spatial and temporal scale. We observed east-to-west spatial genetic variation across the range, which presented evidence of isolation by distance. Different key factors drive adaptation of loblolly pine from different geographical regions. Trees residing near the northeastern edge of the range, spanning across Delaware and Maryland and mountainous areas of Virginia, North Carolina, South Carolina, and northern Georgia, were identified to be most likely impacted by climate change based on the large difference in genetic composition under current and future climate conditions. This study provides new perspectives on adaptive genetic variation of loblolly pine in response to different climate scenarios, and the results can be used to target particular populations while developing adaptive forest management guidelines.

Application of machine learning in safety evaluation of athletes training based on physiological index monitoring

In order to realize the monitoring and adaptive evaluation of athletes' training safety, the physiological index monitoring and training safety evaluation method of athletes based on machine learning is put forward. Taking the physiological index parameters such as heart rate HR, maximum oxygen uptake VO2max, oxygen pulse O2P, respiratory entropy RQ, maximum ventilation (VEmax) as constraint indexes, the physiological index monitoring big data analysis model of athlete training safety evaluation is constructed. The multivariate index joint analysis modeling method is used to reconstruct the characteristics of athletes' physiological index monitoring, the related characteristic quantities of athletes' physiological index monitoring data are extracted, the correlation of athletes' physiological indexes is analyzed by using association rule reconstruction method, and the adaptive training of athletes' physiological index monitoring and safety evaluation is carried out combined with machine learning method. The statistical analysis and optimization control model of athletes' training safety evaluation is constructed, and the physiological index monitoring and training safety evaluation of athletes are realized under machine learning. The simulation results show that the confidence level of this method is high and the convergence of the evaluation process is good, so it has a good application value in athletes training and physiological monitoring.

A Spectral Assignment-Oriented Approach to Improve Interpretability and Accuracy of Proxy Spectral-Based Models

In modeling chemical attributes using hyperspectral data, nonlinear relationships between the predictor and the response are frequent. The common nonlinear modeling techniques improve prediction accuracy but suffer from low interpretability of the models. In this paper, we demonstrate a new multivariate modeling method, denoted as spectral assignment-oriented partial least squares (SAO-PLS), which is designed to provide a nonlinear modeling solution with strong interpretability products. The need for this approach is apparent when a given sample population consists of different spectral features for different levels of the response. Accordingly, the suggested SAO-PLS algorithm segments the data in an optimal location on the response distribution by maximizing the difference in spectral assignments between two clusters. SAO-PLS is applied here to two test cases with different characteristics: 1) an established data set containing airborne hyperspectral data of asphaltic roads, merged with in situ measured dynamic friction values captured using a standardized method and 2) a soil spectral library, spectrally measured with an analytical spectral device spectrometer, to which organic carbon measurements were applied. Our results demonstrate the superiority of SAO-PLS over partial least-squares regression for both model accuracy and interpretability, providing a deeper understanding of the underlying processes.

Grading bunch tightness for grape by multiperspective imaging approach coupled with multivariate classification methods

AbstractBunch's tightness is a significant trait in determining tablegrape and wine quality. But standard determination of tightness needs the trained technicians and complex manual measurements. This study is to validate the feasibility of multiperspective imaging analysis combined with multivariate modeling methods to grade the grape bunch compactness in a rapid, automatic, and objective approach. The detection system was developed by multiperspective projection imaging structure, where mono‐camera was used to simultaneously capture the whole surface of grape bunch by two mirrors' reflections. Through series of image processes, 23 features that described bunch regions, contours, the cluster density, and so forth, were digitalized and normalized for linear projections by principal component analysis (PCA) and partial least square (PLS), respectively. Then the projected latent components were obtained and were taken as input of the developed classifiers, to qualitatively assess the bunch's tightness. Results indicated these components compressed more information from PLS than PCA projection, and PLS‐support vector machine classifier outperformed best with 88.33% accuracy rate. It was concluded that the multiperspective simultaneously imaging approach coupled with image process and pattern recognition technology could be applied to evaluate the external tightness of grape bunch. Moreover, it has the potential to be a good reference for rapid grade of berry cluster.Practical applicationsThe proven performance of machine vision coupled with pattern recognition could be implemented in automatic measurement system to rapidly and objectively classify the tightness of postharvest grape bunch or other berry‐clustered fruit. This work could be a useful application for the specific consumptions to distinguish the tight from the loose on account of that the tightness plays a significant role in determining table grape and wine quality.

Comparative Metabonomic Investigations of Schistosoma japonicum From SCID Mice and BALB/c Mice: Clues to Developmental Abnormality of Schistosome in the Immunodeficient Host.

The growth and development of schistosome has been affected in the immunodeficient hosts. But it remains unresolved about the molecular mechanisms involved in the development and reproduction regulation of schistosomes. This study tested and compared the metabolic profiles of the male and female Schistosoma japonicum worms collected from SCID mice and BALB/c mice at 5 weeks post infection using liquid chromatography tandem mass spectrometry (LC-MS/MS) platform, in which the worms from SCID mice were the investigated organisms and the worms from BALB/c mice were used as the controls. There were 1015 ion features in ESI+ mode and 342 ion features in ESI- mode were identified after filtration by false discovery rate. Distinct metabolic profiles were found to clearly differentiate both male and female worms in SCID mice from those in BALB/c mice using multivariate modeling methods including the Principal Component Analysis (PCA), Partial Least Squares Discriminant Analysis (PLS-DA), and Orthogonal Partial Least Squares Discriminant Analysis (OPLS-DA). There were more differential metabolites in female worms than in male worms between SCID mice and BALB/c mice. And common and uniquely perturbed metabolites and pathways were identified among male and female worms from SCID mice when compared with BALB/c mice. The enriched metabolite sets of the differential metabolites in male worms between SCID mice and BALB/c mice included bile acid biosynthesis, taurine and hypotaurine metabolism, sphingolipid metabolism, retinol metabolism, purine metabolism, etc. And the enriched metabolite sets of differential metabolites in female worms included retinol metabolism, alpha linolenic acid and linoleic acid metabolism, purine metabolism, sphingolipid metabolism, glutamate metabolism, etc. Further detection and comparison in transcript abundance of genes of the perturbed retinol metabolism and its associated meiosis process in worms identified clues suggesting accumulated retinyl ester and perturbed meiotic process. These findings suggested an association between the schistosome with retarded growth and development in SCID mice and their perturbed metabolites and metabolic pathways, and provided a new insight into the growth and development regulation of S. japonicum worms from the metabolic level, which indicated great clues for discovery of drugs or vaccines against the parasites and disease with more researches.

Effective use of statistical tools in agricultural sciences: a critical review of multivariate modelling methods

Objectives: A critical review was conducted to assess effective use of multivariate modelling methods for data analysis in agricultural sciences and related fields. Methodology and Results: Four main agricultural fields were considered: biology; agronomy; ecology; and food nutrition. Two journals were randomly selected per agricultural field and up to 250 articles were downloaded considering a ten-year period (2008-2017) per journal. From papers, information such as: statistical methods used; and whether multivariate modelling methods were required and used for data analysis or not, was recorded. Basic statistical methods: descriptive statistics, univariate parametric tests and related tests (post-hoc tests, normality test, and homoscedasticity tests) were the most frequently used. Advanced statistical methods such as multivariate descriptive and modelling methods, Bayesian methods, recorded the least use values. Multivariate modelling methods were rarely used though they were sometimes required according to agricultural fields. The highest and lowest effective uses of statistical methods were recorded for the agronomy and biology fields, respectively. Conclusion and application of findings: There is a gap between the development of advanced statistical methods, their usefulness and accessibility to analyse data in applied sciences especially agricultural sciences. Further investigations in statistical methods’ development may integrate and justify their usefulness in applied sciences. Collaborations between applied scientists and statisticians are necessary for better analysis of research data. Keywords: multivariate modelling techniques, effective use, agronomy, critical analysis, statistical methods

Classification of Sand Grains by Terahertz Time-Domain Spectroscopy and Chemometrics

A novel method combining terahertz time-domain spectroscopy (THz-TDS) with chemometrics was proposed for classification of sand samples from different deserts and grain sizes. THz-TDS combined with chemometrics was firstly utilized for identifying the source of sand samples and analyzing the sizes of sand granule. Terahertz time-domain transmittance spectra were collected and then Savitzky–Golay smoothing and orthogonal signal correction (OSC) approaches were used for spectral pretreatment. Principal component analysis, partial least squares discriminant analysis, and support vector machine were used to establish classification models of the sand samples. The prediction correctness ranged between 94 and 97%. Scanning electron microscopy was used to investigate the morphology of the sand samples. The sand samples were analyzed by energy dispersive spectroscopy and the elemental content distributions were obtained. Fourier transform infrared absorption spectra of the sand samples were also collected. THz-TDS combined with multivariate modeling methods provided reliable and useful information to determine the provenance of the sands originated from different deserts in northern China. This work reveals that THz-TDS can be a useful tool for classification of sand from different deserts.

Combination of fractional order derivative and memory-based learning algorithm to improve the estimation accuracy of soil organic matter by visible and near-infrared spectroscopy

Visible and near-infrared (Vis–NIR) spectroscopy is used to estimate soil organic matter (SOM). Spectral preprocessing techniques and multivariate modeling methods play important roles in the quantitative analysis of SOM. First and second derivatives (i.e., the conventional integer order derivatives) are commonly used spectral derivatives, which, however, may ignore some detailed spectral information regarding SOM. Here, we presented a fractional order derivative (FOD) method to preprocess the reflectance spectra. Robust modeling methods are still required for accurate estimation of SOM. Local modeling technique (memory-based learning, MBL) was introduced to compare with two global modeling approaches, namely, partial least square (PLS) and random forest (RF). A total of 535 topsoil samples were gathered from Hubei Province, Central China, with their reflectance spectra and SOM contents measured in the laboratory. FOD was allowed to vary from 0 to 2 with an increment of 0.25 at each step. Coefficient of determination (R2) and ratio of the performance to deviation (RPD) were employed as performance statistics during validation. Results showed that with the increase of derivative order, the baseline drifts and overlapping peaks were gradually removed but the spectral strength decreased concurrently. Higher derivative order reflectance (i.e., 1.5-order, 1.75-order, and 2-order reflectance) were more susceptible to spectral noise interferences. The correlation coefficient of SOM with FOD processed spectra at some specific wavelengths was larger than that with the original reflectance. MBL performed better than PLS and RF, regardless of FOD transformation. Calibration with 0.25-order reflectance and MBL provided the most accurate estimation of SOM, with an RPD of 2.23. Our results confirm the effectiveness of FOD and local modeling (MBL) in the development of Vis–NIR models for SOM estimation.

Application of Sparse Linear Discriminant Analysis and Elastic Net for Diagnosis of IgA Nephropathy: Statistical and Biological Viewpoints

Background:IgA nephropathy (IgAN) is the most common primary glomerulonephritis diagnosed based on renal biopsy. Mesangial IgA deposits along with the proliferation of mesangial cells are the histologic hallmark of IgAN. Non-invasive diagnostic tools may help to prompt diagnosis and therapy. The discovery of potential and reliable urinary biomarkers for diagnosis of IgAN depends on applying robust and suitable models. Applying two multivariate modeling methods on a urine proteomic dataset were obtained from IgAN patients, and comparison of the results of these methods were the purpose of this study.Methods:Two models were constructed for urinary protein profiles of 13 patients and 8 healthy individuals, based on sparse linear discriminant analysis (SLDA) and elastic net (EN) regression methods. A panel of selected biomarkers with the best coefficients were proposed and further analyzed for biological relevance using functional annotation and pathway analysis.Results:Transferrin, α1-antitrypsin, and albumin fragments were the most important up-regulated biomarkers, while fibulin-5, YIP1 family member 3, prasoposin, and osteopontin were the most important down-regulated biomarkers. Pathway analysis revealed that complement and coagulation cascades and extracellular matrix-receptor interaction pathways impaired in the pathogenesis of IgAN.Conclusion:SLDA and EN had an equal importance for diagnosis of IgAN and were useful methods for exploring and processing proteomic data. In addition, the suggested biomarkers are reliable candidates for further validation to non-invasive diagnose of IgAN based on urine examination.

Modeling daily suspended sediment load using improved support vector machine model and genetic algorithm

Prediction of sediment volume and sediment load is always one of the important issues for decision-makers of watershed basins. The present study investigated the daily suspended sediment load in a watershed basin using the improved support vector machine method. Since in most of the previous studies, the coefficients of the support vector machine method had been calculated based on trial and error, in the present study, the combination of the support vector machine and the genetic algorithm is used. In the first step, the unknown parameters of the support vector machine are calculated and then, the sediment load simulation is performed. Two case studies in the present work involve two earth dams in Semnan Province called Veynakeh and Royan. Furthermore, multivariate adaptive regression spline (MARS) and MT tree model (M5T) methods are used for comparison. The results indicated that the input combination of discharge data at the current time and one, two, and three previous days has the best performance for all models. Also, the support vector machine-genetic algorithm (SVM-GA) model has a lower root mean square error (RMSE) and mean absolute error (MAE) compared to the MARS and M5T models for both stations. In addition, comparing observational data with simulation data based on the R2 coefficient suggested that the SVM-GA model offers more accurate results than the other two methods. Accordingly, the SVM-GA method used in this study has a high potential for simulating sediment volume.

제한된 상관 데이터의 다변량 통계모델링 기법 비교

제한된 상관 데이터의 다변량 통계모델링 기법 비교

Statins use and risk of new-onset diabetes in hypertensive patients: a population-based retrospective cohort study in Yinzhou district, Ningbo city, People's Republic of China.

BackgroundReports have suggested that statin use is associated with an increased incidence of type 2 diabetes mellitus (T2DM). Guidelines suggested that statins should be prescribed in hypertensive patients for primary prevention. However, there were very few studies on the risk of T2DM associated with statin use among patients with hypertension in mainland People’s Republic of China.PurposeTo determine the association between statin use and new-onset diabetes mellitus among patients with hypertension in mainland People’s Republic of China.Patients and methodsWe performed a retrospective cohort study of hypertensive patients using the Yinzhou regional health care database from January 1, 2010, to August 31, 2016. Patients aged 30–90 years old without T2DM were eligible for inclusion. We identified new statin initiators and nonusers by using prescription records of inpatients and outpatients. Multivariate Cox model and propensity score methods were used to adjust potential confounders, including age, sex, body mass index, comorbidities, lifestyle characteristics, and baseline antihypertensive drug use. The risk of incident T2DM among statin initiators compared to nonusers was estimated by the Cox proportional hazards model. Propensity scores for statin use were then developed using logistic regression, statin initiators were matched 1:1 with nonusers according to propensity scores with the nearest neighbor matching method within 0.2 caliper width, and Cox regression was again conducted.ResultsAmong 67,993 patients (21,551 statin initiators; 46,442 nonusers), the unadjusted incidence rate of incident T2DM was higher in statin initiators than nonusers (25.68 versus 14.19 events/1,000 person-years; adjusted hazard ratio: 1.55; 95% confidence interval: 1.44–1.66). After propensity score 1:1 matching (19,818 statin initiators; 19,818 nonusers), baseline characteristics between 2 groups were balanced except that the nonusers group was 0.53 years older on average (P<0.001). Then statin use was still associated with a significant increased risk for T2DM in the matched cohort (adjusted hazard ratio: 1.54; 95% confidence interval: 1.41–1.67). Subgroup analyses also demonstrated similar findings.ConclusionOur study indicated an association between statin use and an increased risk of new-onset diabetes mellitus. It provides better understanding of statin and new-onset diabetes mellitus association among hypertensive patients in real-word setting. As an observational study, our findings were prone to unmeasured confounding and bias.

A multivariate short-term traffic flow forecasting method based on wavelet analysis and seasonal time series

Short-term traffic flow forecasting is a key step to achieve the performance of intelligent transportation system (ITS). Timely and accurate traffic information prediction is also the prerequisite of realizing proactive traffic control and dynamic traffic assignment effectively. Based on the fact that univariate forecasting methods have limited forecasting abilities when the data is missing or erroneous and that single models make no full use of information underline data, a new hybrid method with multivariate for short-term traffic flow forecasting is proposed. This method combines statistical analysis method with computational intelligence techniques to mine the characteristic of traffic flow as well as forecast short-term traffic state. First, the wavelet de-noising is employed to remove the noise information. Then, time series analysis is used to analyze time-varying and periodic characteristic of traffic flow. Furthermore, the seasonal auto-regressive moving average with external input (SARIMAX) is established to fit traffic flow with occupancy as exogenous variables. Finally, wavelet forecast is adopted to forecast the values of occupancy which are used as exogenous input, and a WSARIMAX is constructed to forecast traffic flow. Using the relationship of flow and occupancy at the same road section and taking traffic flow and occupancy data from freeway I-694 EB in the Twin Cities as endogenous variables and exogenous variables respectively, this paper studies the forecasting performance of the proposed method. The study results are encouraging. Compared with SARIMA newly proposed in literature, WSARIMA and SARIMAX improved method with wavelet analysis and multivariate modeling method, the proposed method gets improvements of 12.95%, 12.62% and 10.41% in forecasting accuracy of one-step ahead respectively. For ten-steps ahead forecasting, it gets improvements of 18.87%, 17.05% and 2.57% in forecasting accuracy respectively.