Machine learning-based nonlinear regression-adjusted real-time quality control modeling: a multi-center study

OBJECTIVES

Patient-based real-time quality control (PBRTQC), a laboratory tool for monitoring the performance of the testing process, has gained increasing attention in recent years. It has been questioned for its generalizability among analytes, instruments, laboratories, and hospitals in real-world settings. Our purpose was to build a machine learning, nonlinear regression-adjusted, patient-based real-time quality control (mNL-PBRTQC) with wide application.

METHODS

Using computer simulation, artificial biases were added to patient population data of 10 measurands. An mNL-PBRTQC was created using eight hospital laboratory databases as a training set and validated by three other hospitals' independent patient datasets. Three different Patient-based models were compared on these datasets, the IFCC PBRTQC model, linear regression-adjusted real-time quality control (L-RARTQC), and the mNL-PBRTQC model.

RESULTS

Our study showed that in the three independent test data sets, mNL-PBRTQC outperformed the IFCC PBRTQC and L-RARTQC for all measurands and all biases. Using platelets as an example, it was found that for 20 % bias, both positive and negative, the uncertainty of error detection for mNL-PBRTQC was smallest at the median and maximum values.

CONCLUSIONS

mNL-PBRTQC is a robust machine learning framework, allowing accurate error detection, especially for analytes that demonstrate instability and for detecting small biases.

A Deep Learning Approach to Analyzing Continuous-Time Cognitive Processes

The dynamics of the mind are complex. Mental processes unfold continuously in time and may be sensitive to a myriad of interacting variables, especially in naturalistic settings. But statistical models used to analyze data from cognitive experiments often assume simplistic dynamics. Recent advances in deep learning have yielded startling improvements to simulations of dynamical cognitive processes, including speech comprehension, visual perception, and goal-directed behavior. But due to poor interpretability, deep learning is generally not used for scientific analysis. Here, we bridge this gap by showing that deep learning can be used, not just to imitate, but to analyze complex processes, providing flexible function approximation while preserving interpretability. To do so, we define and implement a nonlinear regression model in which the probability distribution over the response variable is parameterized by convolving the history of predictors over time using an artificial neural network, thereby allowing the shape and continuous temporal extent of effects to be inferred directly from time series data. Our approach relaxes standard simplifying assumptions (e.g., linearity, stationarity, and homoscedasticity) that are implausible for many cognitive processes and may critically affect the interpretation of data. We demonstrate substantial improvements on behavioral and neuroimaging data from the language processing domain, and we show that our model enables discovery of novel patterns in exploratory analyses, controls for diverse confounds in confirmatory analyses, and opens up research questions in cognitive (neuro)science that are otherwise hard to study.

Large-scale evidence for logarithmic effects of word predictability on reading time

During real-time language comprehension, our minds rapidly decode complex meanings from sequences of words. The difficulty of doing so is known to be related to words' contextual predictability, but what cognitive processes do these predictability effects reflect? In one view, predictability effects reflect facilitation due to anticipatory processing of words that are predictable from context. This view predicts a linear effect of predictability on processing demand. In another view, predictability effects reflect the costs of probabilistic inference over sentence interpretations. This view predicts either a logarithmic or a superlogarithmic effect of predictability on processing demand, depending on whether it assumes pressures toward a uniform distribution of information over time. The empirical record is currently mixed. Here, we revisit this question at scale: We analyze six reading datasets, estimate next-word probabilities with diverse statistical language models, and model reading times using recent advances in nonlinear regression. Results support a logarithmic effect of word predictability on processing difficulty, which favors probabilistic inference as a key component of human language processing.

The Implementation of Neural Networks for Polymer Mold Surface Evaluation

This paper presents the measurement and evaluation of the surfaces of molds produced using additive technologies. This is an emerging trend in mold production. The surfaces of such molds must be treated, usually using laser-based alternative machining methods. Regular evaluation is necessary because of the gradually deteriorating quality of the mold surface. However, owing to the difficulty in scanning the original surface of the injection mold, it is necessary to perform surface replication. Therefore, this study aims to describe the production of surface replicas for in-house developed polymer molds together with the determination of suitable descriptive parameters, the method of comparing variances, and the mean values for the surface evaluation. Overall, this study presents a new summary of the evaluation process of replicas of the surfaces of polymer molds. The nonlinear regression methodology provides the corresponding functional dependencies between the relevant parameters. The statistical significance of a neural network with two hidden layers based on the principle of Rosenblatt's perceptron has been proposed and verified. Additionally, machine learning was utilized to better compare the original surface and its replica.

Experimental and computational approach to establish fit-for-purpose cell viability assays

Aim: Cell viability assays are critical for cell-based products. Here, we demonstrate a combined experimental and computational approach to identify fit-for-purpose cell assays that can predict changes in cell proliferation, a critical biological response in cell expansion. Materials & methods: Jurkat cells were systematically injured using heat (45 ± 1°C). Cell viability was measured at 0 h and 24 h after treatment using assays for membrane integrity, metabolic function and apoptosis. Proliferation kinetics for longer term cultures were modeled using the Gompertz distribution to establish predictive models between cell viability results and proliferation. Results & conclusion: We demonstrate an approach for ranking these assays as predictors of cell proliferation and for setting cell viability specifications when a particular proliferation response is required.

A comparative study of in vitro dose-response estimation under extreme observations

Quantifying drug potency, which requires an accurate estimation of dose-response relationship, is essential for drug development in biomedical research and life sciences. However, the standard estimation procedure of the median-effect equation to describe the dose-response curve is vulnerable to extreme observations in common experimental data. To facilitate appropriate statistical inference, many powerful estimation tools have been developed in R, including various dose-response packages based on the nonlinear least squares method with different optimization strategies. Recently, beta regression-based methods have also been introduced in estimation of the median-effect equation. In theory, they can overcome nonnormality, heteroscedasticity, and asymmetry and accommodate flexible robust frameworks and coefficients penalization. To identify a reliable estimation method(s) to estimate dose-response curves even with extreme observations, we conducted a comparative study to review 14 different tools in R and examine their robustness and efficiency via Monte Carlo simulation under a list of comprehensive scenarios. The simulation results demonstrate that penalized beta regression using the mgcv package outperforms other methods in terms of stable, accurate estimation, and reliable uncertainty quantification.

A Robust Noise Estimation Algorithm Based on Redundant Prediction and Local Statistics

Blind noise level estimation is a key issue in image processing applications that helps improve the visualization and perceptual quality of images. In this paper, we propose an improved block-based noise level estimation algorithm. The proposed algorithm first extracts homogenous patches from a single noisy image using local features, obtaining the covariance matrix eigenvalues of the patches, and constructs dynamic thresholds for outlier discrimination. By analyzing the correlations between scene complexity, noise strength, and other parameters, a nonlinear discriminant coefficient regression model is fitted to accurately predict the number of redundant dimensions and calculate the actual noise level according to the statistical properties of the elements in the redundancy dimension. The experimental results show that the accuracy and robustness of the proposed algorithm are better than those of the existing noise estimation algorithms in various scenes under different noise levels. It performs well overall in terms of performance and execution speed.

Nonlinear function-on-scalar regression via functional universal approximation

We consider general nonlinear function-on-scalar (FOS) regression models, where the functional response depends on multiple scalar predictors in a general unknown nonlinear form. Existing methods either assume specific model forms (e.g., additive models) or directly estimate the nonlinear function in a space with dimension equal to the number of scalar predictors, which can only be applied to models with a few scalar predictors. To overcome these shortcomings, motivated by the classic universal approximation theorem used in neural networks, we develop a functional universal approximation theorem which can be used to approximate general nonlinear FOS maps and can be easily adopted into the framework of functional data analysis. With this theorem and utilizing smoothness regularity, we develop a novel method to fit the general nonlinear FOS regression model and make predictions. Our new method does not make any specific assumption on the model forms, and it avoids the direct estimation of nonlinear functions in a space with dimension equal to the number of scalar predictors. By estimating a sequence of bivariate functions, our method can be applied to models with a relatively large number of scalar predictors. The good performance of the proposed method is demonstrated by empirical studies on various simulated and real datasets.

Machine Learning Techniques for the Performance Enhancement of Multiple Classifiers in the Detection of Cardiovascular Disease from PPG Signals

Photoplethysmography (PPG) signals are widely used in clinical practice as a diagnostic tool since PPG is noninvasive and inexpensive. In this article, machine learning techniques were used to improve the performance of classifiers for the detection of cardiovascular disease (CVD) from PPG signals. PPG signals occupy a large amount of memory and, hence, the signals were dimensionally reduced in the initial stage. A total of 41 subjects from the Capno database were analyzed in this study, including 20 CVD cases and 21 normal subjects. PPG signals are sampled at 200 samples per second. Therefore, 144,000 samples per patient are available. Now, a one-second-long PPG signal is considered a segment. There are 720 PPG segments per patient. For a total of 41 subjects, 29,520 segments of PPG signals are analyzed in this study. Five dimensionality reduction techniques, such as heuristic- (ABC-PSO, cuckoo clusters, and dragonfly clusters) and transformation-based techniques (Hilbert transform and nonlinear regression) were used in this research. Twelve different classifiers, such as PCA, EM, logistic regression, GMM, BLDC, firefly clusters, harmonic search, detrend fluctuation analysis, PAC Bayesian learning, KNN-PAC Bayesian, softmax discriminant classifier, and detrend with SDC were utilized to detect CVD from dimensionally reduced PPG signals. The performance of the classifiers was assessed based on their metrics, such as accuracy, performance index, error rate, and a good detection rate. The Hilbert transform techniques with the harmonic search classifier outperformed all other classifiers, with an accuracy of 98.31% and a good detection rate of 96.55%.

Introduction to the Use of Linear and Nonlinear Regression Analysis in Quantitative Biological Assays

Biological assays are essential tools in biomedical and pharmaceutical research. In simplest terms, such an assay is an analytical method used to measure or predict a response in a biological system in the presence of a given stimulus (e.g., drug). The inherent complexity involved in evaluating a biological system requires the use of rigorous and appropriate tools for data analysis. Linear and nonlinear regression models represent critically important statistical analyses used to define the relationships between variables of interest in biological systems. Recent challenges relating to the reproducibility of published data suggest the absence of standardized and routine use of statistics to support experimental results across a wide range of scientific disciplines. The current situation warrants an introductory review of basic regression concepts using current, practical examples, along with references to in-depth resources. The goal is to provide the necessary information to help standardize the analysis of biological assays in academic research and drug discovery and development, elevating their utility and increasing data transparency and reproducibility. © 2023 The Authors. Current Protocols published by Wiley Periodicals LLC.

Variable selection in nonlinear function-on-scalar regression

We develop a new method for variable selection in a nonlinear additive function-on-scalar regression (FOSR) model. Existing methods for variable selection in FOSR have focused on the linear effects of scalar predictors, which can be a restrictive assumption in the presence of multiple continuously measured covariates. We propose a computationally efficient approach for variable selection in existing linear FOSR using functional principal component scores of the functional response and extend this framework to a nonlinear additive function-on-scalar model. The proposed method provides a unified and flexible framework for variable selection in FOSR, allowing nonlinear effects of the covariates. Numerical analysis using simulation study illustrates the advantages of the proposed method over existing variable selection methods in FOSR even when the underlying covariate effects are all linear. The proposed procedure is demonstrated on accelerometer data from the 2003-2004 cohorts of the National Health and Nutrition Examination Survey (NHANES) in understanding the association between diurnal patterns of physical activity and demographic, lifestyle, and health characteristics of the participants.

Bayesian additive regression trees for multivariate skewed responses

This paper introduces a nonparametric regression approach for univariate and multivariate skewed responses using Bayesian additive regression trees (BART). Existing BART methods use ensembles of decision trees to model a mean function, and have become popular recently due to their high prediction accuracy and ease of use. The usual assumption of a univariate Gaussian error distribution, however, is restrictive in many biomedical applications. Motivated by an oral health study, we provide a useful extension of BART, the skewBART model, to address this problem. We then extend skewBART to allow for multivariate responses, with information shared across the decision trees associated with different responses within the same subject. The methodology accommodates within-subject association, and allows varying skewness parameters for the varying multivariate responses. We illustrate the benefits of our multivariate skewBART proposal over existing alternatives via simulation studies and application to the oral health dataset with bivariate highly skewed responses. Our methodology is implementable via the R package skewBART, available on GitHub.

Biomass allometric models for Larix rupprechtii based on Kosak's taper curve equations and nonlinear seemingly unrelated regression

The diameter at breast height (DBH) is the most important independent variable in biomass allometry models based on metabolic scaling theory (MST) or geometric theory. However, the fixed position DBH can be misleading in its use of universal scaling laws and lead to some deviation for the biomass model. Therefore, it is still an urgent scientific problem to build a high-precision biomass model system. A dataset of 114 trees was destructively sampled to obtain dry biomass components, including stems, branches, and foliage, and taper measurements to explore the applicability of biomass components to allometric scaling laws and develop a new system of additive models with the diameter in relative height (DRH) for each component of a Larch (Larix principis-rupprechtii Mayr) plantation in northern China. The variable exponential taper equations were modelled using nonlinear regression. In addition, applying nonlinear regression and nonlinear seemingly unrelated regression (NSUR) enabled the development of biomass allometric models and the system of additive models with DRH for each component. The results showed that the Kozak's (II) 2004 variable exponential taper equation could accurately describe the stem shape and diameter in any height of stem. When the diameters in relative height were D_0.2, D_0.5, and D_0.5 for branches, stems, and foliage, respectively, the allometric exponent of the stems and branches was the closest to the scaling relations predicted by the MST, and the allometric exponent of foliage was the most closely related to the scaling relations predicted by geometry theory. Compared with the nonlinear regression, the parameters of biomass components estimated by NSUR were lower, and it was close to the theoretical value and the most precise at forecasting. In the study of biomass process modelling, utilizing the DRH by a variable exponential taper equation can confirm the general biological significance more than the DBH of a fixed position.

Modeling net energy partition patterns of growing-finishing pigs using nonlinear regression and artificial neural networks

The objectives of this study were to evaluate the net energy (NE) partition patterns of growing-finishing pigs at different growing stages and to develop the corresponding prediction models using nonlinear regression (NLR) and artificial neural networks (ANN). Twenty-four pigs with an initial body weight (BW) of ~30 kg were kept in metabolic cages and fed ad libitum and were moved into six respiration chambers in turns until ~90 kg. The NE partition patterns, i.e., NE for maintenance (NEm), NE retained as protein (NEp), and NE retained as lipid (NEl), were calculated based on indirect calorimetry and nitrogen balance techniques. The energy balance data collected through the animal trial was then randomly split into a training data set containing 75% of the samples and a testing data set containing the remaining 25% of the samples. The NLR models and a series of ANN models were established on the training data set to predict the metabolizable energy intake, NE intake, NEm, NEp, and NEl of pigs. The best-fitted ANN models were selected by 5-fold cross-validation in the training data set. The prediction performance of the best-fitted NLR and ANN models were compared on the testing data set. The results showed that the average NE intakes of pigs were 17.71, 23.25, 24.56, and 28.96 MJ/d in 30 to 45 kg, 45 to 60 kg, 60 to 75 kg, and 75 to 90 kg, respectively. The NEm and NEl (MJ/d) kept increasing as BW increased from 30 kg to 90 kg, while the NEp increased to its maximum value and then kept in a certain range of 4.64 to 4.88 MJ/d. The proportion of NEm for pigs at 30 to 90 kg stayed within the range of 42.0% to 48.6%, while the proportion of NEl kept increasing. For the prediction models built based on the animal trial, ANN models exhibited better performance than NLR models for all the target outputs. In conclusion, NE partition patterns changed in different growth stages of pigs, and ANN models are more flexible and powerful than NLR models in predicting the NE partition patterns of growing-finishing pigs.

Inappropriate Evaluation of Effect Modifications Based on Categorical Outcomes: A Systematic Review of Randomized Controlled Trials

Our meta-epidemiological study aimed to describe the prevalence of reporting effect modification only on relative scale outcomes and inappropriate interpretations of the coefficient of interaction terms in nonlinear models on categorical outcomes. Our study targeted articles published in the top 10 high-impact-factor journals between 1 January and 31 December 2021. We included two-arm, parallel-group, interventional superiority randomized controlled trials to evaluate the effects of modifications on categorical outcomes. The primary outcomes were the prevalence of reporting effect modifications only on relative scale outcomes and that of inappropriately interpreting the coefficient of interaction terms in nonlinear models on categorical outcomes. We included 52 articles, of which 41 (79%) used nonlinear regression to evaluate effect modifications. At least 45/52 articles (87%) reported effect modifications based only on relative scale outcomes, and at least 39/41 (95%) articles inappropriately interpreted the coefficient of interaction terms merely as indices of effect modifications. The quality of the evaluations of effect modifications in nonlinear models on categorical outcomes was relatively low, even in randomized controlled trials published in medical journals with high impact factors. Researchers should report effect modifications of both absolute and relative scale outcomes and avoid interpreting the coefficient of interaction terms in nonlinear regression analyses.

Toward Precision Radiotherapy: A Nonlinear Optimization Framework and an Accelerated Machine Learning Algorithm for the Deconvolution of Tumor-Infiltrating Immune Cells

Tumor-infiltrating immune cells (TIICs) form a critical part of the ecosystem surrounding a cancerous tumor. Recent advances in radiobiology have shown that, in addition to damaging cancerous cells, radiotherapy drives the upregulation of immunosuppressive and immunostimulatory TIICs, which in turn impacts treatment response. Quantifying TIICs in tumor samples could form an important predictive biomarker guiding patient stratification and the design of radiotherapy regimens and combined immune-radiation treatments. As a result of several limitations associated with experimental methods for quantifying TIICs and the availability of extensive gene sequencing data, deconvolution-based computational methods have appeared as a suitable alternative for quantifying TIICs. Accordingly, we introduce and discuss a nonlinear regression approach (remarkably different from the traditional linear modeling approach of current deconvolution-based methods) and a machine learning algorithm for approximating the solution of the resulting constrained optimization problem. This way, the deconvolution problem is treated naturally, given that the gene expression levels of pure and heterogenous samples do not have a strictly linear relationship. When applied across transcriptomics datasets, our approach, which also allows the coupling of different loss functions, yields results that closely match ground-truth values from experimental methods and exhibits superior performance over popular deconvolution-based methods.

Nonlinear Tactile Estimation Model Based on Perceptibility of Mechanoreceptors Improves Quantitative Tactile Sensing

Tactile sensing has attracted significant attention as a tactile quantitative evaluation method because the tactile sensation is an important factor while evaluating consumer products. Although the human tactile perception mechanism has nonlinearity, previous studies have often developed linear regression models. In contrast, this study proposes a nonlinear tactile estimation model that can estimate sensory evaluation scores from physical measurements. We extracted features from the vibration data obtained by a tactile sensor based on the perceptibility of mechanoreceptors. In parallel, a sensory evaluation test was conducted using 10 evaluation words. Then, the relationship between the extracted features and the tactile evaluation results was modeled using linear/nonlinear regressions. The best model was concluded by comparing the mean squared error between the model predictions and the actual values. The results imply that there are multiple evaluation words suitable for adopting nonlinear regression models, and the average error was 43.8% smaller than that of building only linear regression models.

A New Creep-Fatigue Interaction Model for Predicting Deformation of Coarse-Grained Soil

Studying the creep-fatigue interaction of the coarse-grained soil (CGS) is very important for safety assessment and disaster prevention in subgrade engineering. Current research work is mainly focused on single creep or fatigue deformation. In this paper, a new creep-fatigue interaction model is established to predict the creep-fatigue interaction deformation of different gradation CGS based on the rheological mechanics and the interactive relationship between creep and fatigue complex compliance method. Triaxial creep-fatigue interaction tests of different gradations CGS under different average stresses and frequencies were conducted to verify the new creep-fatigue interaction model. Research results show that for the creep-fatigue and fatigue-creep interaction, the fatigue deformation is always larger than the creep deformation under the same stress level. For the creep-fatigue multi-interaction, the second creep and fatigue deformation are always smaller than the first creep and fatigue deformation. The results of the triaxial creep-fatigue interaction tests verify the validity of this new model.

Validation and Improvement of COCTS/HY-1C Sea Surface Temperature Products

In oceanographic study, satellite-based sea surface temperature (SST) retrieval has always been the focus of researchers. This paper investigates several multi-channel SST retrieval algorithms for the thermal infrared band, and evaluates the accuracy of the COCTS/HY-1C SST products. NEAR-GOOS in situ SST data are utilized for validation and improvement, and a three-step matching procedure including geographic location screening, cloud masking, and homogeneity check is conducted to match in situ SST data with satellite SST data. Two improvement schemes, including nonlinear regression and regularization iteration, are proposed to improve the accuracy of the COCTS/HY-1C SST products and the typical application scenarios and the algorithm characteristics of these two schemes are discussed. The standard deviation of residual between retrieved SST and measured SST for these two data improvement algorithms, which are considered as the main indexes for assessment, result in an improvement of 13.245% and 14.096%, respectively. In addition, the generalization ability of the SST models under two data improvement methods is quantitatively compared, and the factors affecting the model accuracy are also carefully evaluated, including the in situ data acquisition method and measurement time (day/night). Finally, future works about SST retrieval with COCTS/HY-1C satellite data are summarized.

Preliminary Eye-Tracking Data as a Nonintrusive Marker for Blood Δ-9-Tetrahydrocannabinol Concentration and Drugged Driving

Background: Cannabis is one of the drugs most often found in drivers involved in serious motor vehicle collisions. Validity and reliability of roadside cannabis detection strategies are questioned. This pilot study aimed to investigate the relationship between eye characteristics and cannabis effects in simulated driving to inform potential development of an alternative detection strategy. Materials and Methods: Multimodal data, including blood samples, eye-tracking recordings, and driving performance data, were acquired from 10 participants during a prolonged single-session driving simulator experiment. The study session included a baseline driving trial before cannabis exposure and seven trials at various times over ∼5 h after exposure. The multidimensional eye-tracking recording from each driving trial for each participant was segmented into nonoverlapping epochs (time windows); 34 features were extracted from each epoch. Blood Δ-9-tetrahydrocannabinol (THC) concentration, standard deviation of lateral position (SDLP), and mean vehicle speed were target variables. The cross-correlation between the temporal profile of each eye-tracking feature and target variable was assessed and a nonlinear regression analysis evaluated temporal trend of features following cannabis exposure. Results: Mean pupil diameter (r=0.81-0.86) and gaze pitch angle standard deviation (r=0.79-0.87) were significantly correlated with blood THC concentration (p<0.01) for all epoch lengths. For driving performance variables, saccade-related features were among those showing the most significant correlation (r=0.61-0.83, p<0.05). Epoch length significantly affected correlations between eye-tracking features and speed (p<0.05), but not SDLP or blood THC concentration (p>0.1). Temporal trend analysis of eye-tracking features after cannabis also showed a significant increasing trend (p<0.01) in saccade-related features, including velocity, scanpath, and duration, as the influence of cannabis decreased by time. A decreasing trend was observed for fixation percentage and mean pupil diameter. Due to the lack of placebo control in this study, these results are considered preliminary. Conclusion: Specific eye characteristics could potentially be used as nonintrusive markers of THC presence and driving-related effects of cannabis. clinicaltrials.gov (NCT03813602).

Microrisk Lab: An Online Freeware for Predictive Microbiology

Microrisk Lab is an R-based online modeling freeware designed to realize parameter estimation and model simulation in predictive microbiology. A total of 36 peer-reviewed models were integrated for parameter estimation (including primary models of bacterial growth/inactivation under static and nonisothermal conditions, secondary models of specific growth rate, and competition models of two-flora growth) and model simulation (including integrated models of deterministic or stochastic bacterial growth/inactivation under static and nonisothermal conditions) in Microrisk Lab. Each modeling section was designed to provide numerical and graphical results with comprehensive statistical indicators depending on the appropriate data set and/or parameter setting. In this study, six case studies were reproduced in Microrisk Lab and compared in parallel with DMFit, GInaFiT, IPMP 2013/GraphPad Prism, Bioinactivation FE, and @Risk, respectively. The estimated and simulated results demonstrated that the performance of Microrisk Lab was statistically equivalent to that of other existing modeling systems. Microrisk Lab allows for a friendly user experience when modeling microbial behaviors owing to its interactive interfaces, high integration, and interconnectivity. Users can freely access this application at https://microrisklab.shinyapps.io/english/ or https://microrisklab.shinyapps.io/chinese/.

Predictive Model for the Surface Tension Changes of Chemical Solutions Used in a Clean-in-Place System

The paper presents the results concerning the influence of concentration and storage time on the equilibrium surface tension of chemical solutions used in a clean-in place (CIP) system. Standard cleaning solutions (prepared under laboratory conditions) and industrial solutions (used in a CIP system in a brewery) were subjected to tests. Solutions from the brewery were collected after being regenerated and changes in equilibrium surface tension were studied during a three-month storage. In the statistical analysis of the solutions, standard deviations were determined in relation to the averages, and a Tukey’s multiple comparison test was performed to determine the effect of dependent variables on the surface tension of solutions. From the results, a nonlinear regression model was developed that provided a mathematical description of the kinetics of changes in the wetting properties of the solutions during their storage. A linear–logarithmic function was adopted to describe the regeneration. Numerical calculations were performed based on the nonlinear least squares method using the Gauss–Newton algorithm. The adequacy of the regression models with respect to the empirical data was verified by the coefficient of determination R and the standard error of estimation Se. The results showed that as the concentration of the substance in the cleaning solution increased, its wetting properties decreased. The same effect was observed with increased storage time as the greatest changes occurred during the first eight weeks. The study also showed that the use of substances to stabilize the cleaning solutions prevented deterioration of their wetting properties, regardless of the concentration of the active substance or storage time.

Fast Grid Search and Bootstrap-based Inference for Continuous Two-phase Polynomial Regression Models

Two-phase polynomial regression models (Robison, 1964; Fuller, 1969; Gallant and Fuller, 1973; Zhan et al., 1996) are widely used in ecology, public health, and other applied fields to model nonlinear relationships. These models are characterized by the presence of threshold parameters, across which the mean functions are allowed to change. That the threshold is a parameter of the model to be estimated from the data is an essential feature of two-phase models. It distinguishes them, and more generally, multi-phase models, from the spline models and has profound implications for both computation and inference for the models. Estimation of two-phase polynomial regression models is a non-convex, non-smooth optimization problem. Grid search provides high quality solutions to the estimation problem, but is very slow when done by brute force. Building upon our previous work on piecewise linear two-phase regression models estimation, we develop fast grid search algorithms for two-phase polynomial regression models and demonstrate their performance. Furthermore, we develop bootstrap-based pointwise and simultaneous confidence bands for mean functions. Monte Carlo studies are conducted to demonstrate the computational and statistical properties of the proposed methods. Three real datasets are used to help illustrate the application of two-phase models, with special attention on model choice.

Can China's Environmental Regulations Effectively Reduce Pollution Emissions?

Based on the provincial panel data of China during 2006–2017, this study uses the panel smooth transition (PSTR) model to study the dynamic transformation mechanism of pollution emission under environmental regulation. We focus on technological progress, economic growth, and foreign direct investment (FDI) as threshold variables, and analyses the non-linear effects of environmental regulation on pollution emissions under those threshold variables, attempting to explore the effectiveness of existing environmental regulations. The structure of biased technological progress is based on the slacks-based measure (SBM) and Global-Malmquist–Luenberger index, which is divided into pollution-biased technology progress and clean-biased technology progress. Finally, we use the panel vector auto regressive (PVAR) algorithm to further verify the relationship. The findings are as follows: (1) Environmental regulation has a significant nonlinear effect on pollution emissions, and technological progress is the optimal threshold variable of this study. (2) Under the influence of these three factors, environmental regulation has a substitution effect on pollution discharge, and a stronger substitution effect on emission reduction in areas with advanced technology and high FDI. It also has a lower emission reduction effect in the high-system areas of economic development than in the low-system areas. (3) The PVAR results show that the impact on environmental regulation of technological progress and FDI has gradually turned from positive to negative; the impact of economic growth on environmental regulation has always been positive but is gradually decreasing. This study points out the direction for governments and companies to implement effective environmental regulations.

Effective Automatic Method Selection for Nonlinear Regression Modeling

Metalearning, an important part of artificial intelligence, represents a promising approach for the task of automatic selection of appropriate methods or algorithms. This paper is interested in recommending a suitable estimator for nonlinear regression modeling, particularly in recommending either the standard nonlinear least squares estimator or one of such available alternative estimators, which is highly robust with respect to the presence of outliers in the data. The authors hold the opinion that theoretical considerations will never be able to formulate such recommendations for the nonlinear regression context. Instead, metalearning is explored here as an original approach suitable for this task. In this paper, four different approaches for automatic method selection for nonlinear regression are proposed and computations over a training database of 643 real publicly available datasets are performed. Particularly, while the metalearning results may be harmed by the imbalanced number of groups, an effective approach yields much improved results, performing a novel combination of supervised feature selection by random forest and oversampling by synthetic minority oversampling technique (SMOTE). As a by-product, the computations bring arguments in favor of the very recent nonlinear least weighted squares estimator, which turns out to outperform other (and much more renowned) estimators in a quite large percentage of datasets.

Probing Substrate/Catalyst Effects Using QSPR Analysis on Friedel-Crafts Acylation Reactions over Hierarchical BEA Zeolites

Attempts to optimize heterogeneous catalysis often lack quantitative comparative analysis. The use of kinetic modelling leads to rate (k) and relative sorption equilibrium constants (K), which can be further rationalized using Quantitative Structure-Property Relationships (QSPR) based on Multiple Linear Regressions (MLR). Friedel-Crafts acylation using commercial and hierarchical BEA zeolites as heterogeneous catalysts, acetic anhydride as the acylating agent, and a set of seven substrates with different sizes and chemical functionalities were herein studied. Catalytic results were correlated with the physicochemical properties of substrates and catalysts. From this analysis, a robust set of equations was obtained allowing inferences about the dominant factors governing the processes. Not entirely surprising, the rate and sorption equilibrium constants were found to be explained in part by common factors but of opposite signs: higher and stronger adsorption forces increase reaction rates, but they also make the zeolite active sites less accessible to new reactant molecules. The most relevant parameters are related to the substrates’ molecular size, which can be associated with different reaction steps, namely accessibility to micropores, diffusion capacity, and polarizability of molecules. The relatively large set of substrates used here reinforces previous findings and brings further insights into the factors that hamper/speed up Friedel-Crafts reactions in heterogeneous media.

Allometric growth in mass by the brain of mammals

I re-examined published data for ontogenetic change in relative mass of the brain in six species of mammal (i.e., sheep, pig, cow, horse, rat, cat) to illustrate an insidious problem with conventional analyses of brain-body allometry. Graphical displays of logarithmic transformations of the original data for each species give the appearance of two discrete mathematical distributions, but untransformed observations nonetheless conform to a single distribution that is well described by a single, nonlinear equation. The concept of biphasic, allometric growth by the brain consequently is an artifact of transformation. The notion of Rapid and Slow phases in relative growth by the brain also is an artifact, because the notion is based explicitly on the concept of biphasic growth allometry. Relative growth by the brain in sheep, pigs, cows, and horses follows the path of a power curve with an exponent less than 1, so relative growth declines progressively as animals grow to their maximum size, at which point growth effectively ends for both brain and body. Relative growth by the brain in rats and cats follows the path of an exponential curve and consequently is more like relative growth by the brain of odontocoete cetaceans and primates, with the brain growing rapidly relative to the body early in ontogeny and attaining maximum (cats) or near-maximum (rats) mass well before the body reaches its maximum. An exponential pattern of relative growth by the brain appears to have evolved independently in rodents, carnivores, odontocoetes, and primates.

Radar Cross Section Near-Field to Far-Field Prediction for Isotropic-Point Scattering Target Based on Regression Estimation

Radar cross section near-field to far-field transformation (NFFFT) is a well-established methodology. Due to the testing range constraints, the measured data are mostly near-field. Existing methods employ electromagnetic theory to transform near-field data into the far-field radar cross section, which is time-consuming in data processing. This paper proposes a flexible framework, named Neural Networks Near-Field to Far-Filed Transformation (NN-NFFFT). Unlike the conventional fixed-parameter model, the near-field RCS to far-field RCS transformation process is viewed as a nonlinear regression problem that can be solved by our fast and flexible neural network. The framework includes three stages: Near-Field and Far-field dataset generation, regression estimator training, and far-field data prediction. In our framework, the Radar cross section prior information is incorporated in the Near-Field and Far-field dataset generated by a group of point-scattering targets. A lightweight neural network is then used as a regression estimator to predict the far-field RCS from the near-field RCS observation. For the target with a small RCS, the proposed method also has less data acquisition time. Numerical examples and extensive experiments demonstrate that the proposed method can take less processing time to achieve comparable accuracy. Besides, the proposed framework can employ prior information about the real scenario to improve performance further.

On Nonlinear Regression for Trends in Split-Belt Treadmill Training

Single and double exponential models fitted to step length symmetry series are used to evaluate the timecourse of adaptation and de-adaptation in instrumented split-belt treadmill tasks. Whilst the nonlinear regression literature has developed substantially over time, the split-belt treadmill training literature has not been fully utilising the fruits of these developments. In this research area, the current methods of model fitting and evaluation have three significant limitations: (i) optimisation algorithms that are used for model fitting require a good initial guess for regression parameters; (ii) the coefficient of determination (R2) is used for comparing and evaluating models, yet it is considered to be an inadequate measure of fit for nonlinear regression; and, (iii) inference is based on comparison of the confidence intervals for the regression parameters that are obtained under the untested assumption that the nonlinear model has a good linear approximation. In this research, we propose a transformed set of parameters with a common language interpretation that is relevant to split-belt treadmill training for both the single and double exponential models. We propose parameter bounds for the exponential models which allow the use of particle swarm optimisation for model fitting without an initial guess for the regression parameters. For model evaluation and comparison, we propose the use of residual plots and Akaike's information criterion (AIC). A method for obtaining confidence intervals that does not require the assumption of a good linear approximation is also suggested. A set of MATLAB (MathWorks, Inc., Natick, MA, USA) functions developed in order to apply these methods are also presented. Single and double exponential models are fitted to both the group-averaged and participant step length symmetry series in an experimental dataset generating new insights into split-belt treadmill training. The proposed methods may be useful for research involving analysis of gait symmetry with instrumented split-belt treadmills. Moreover, the demonstration of the suggested statistical methods on an experimental dataset may help the uptake of these methods by a wider community of researchers that are interested in timecourse of motor training.

Grain Growth Kinetics of 0.65Ca0.61La0.26TiO3-0.35Sm(Mg0.5Ti0.5)O3 Dielectric Ceramic

The 0.65Ca_0.61La_0.26TiO₃-0.35Sm(Mg_0.5Ti_0.5)O₃[0.65CLT-0.35SMT] ceramic was prepared by the solid-state reaction method. The effects of sintering process on its microstructure and grain growth behavior were investigated. The Hillert model and a simplified Sellars model were established by linear regression, and the Sellars-Anelli model with a time index was established by using a nonlinear regression method. The results show that the grain size gradually increases with the increase of sintering temperature and holding time. Meanwhile, the sintering temperature has a more significant effect on the grain growth. The grain sizes of 0.65CLT-0.35SMT ceramic were predicted by the three models and compared with the experimentally measured grain size. The results indicate that for the 0.65CLT-0.35SMT ceramic, the Hillert model has the lowest prediction accuracy and the Sellars-Anelli model, the highest prediction accuracy. In this work, the Sellars-Anelli model can effectively predict the grain growth process of 0.65CLT-0.35SMT ceramic.

A Comparative Study of Biomechanical and Geometrical Attributes of Abdominal Aortic Aneurysms in the Asian and Caucasian Populations

In this work, we provide a quantitative assessment of the biomechanical and geometric features that characterize abdominal aortic aneurysm (AAA) models generated from 19 Asian and 19 Caucasian diameter-matched AAA patients. 3D patient-specific finite element models were generated and used to compute peak wall stress (PWS), 99th percentile wall stress (99th WS), and spatially averaged wall stress (AWS) for each AAA. In addition, 51 global geometric indices were calculated, which quantify the wall thickness, shape, and curvature of each AAA. The indices were correlated with 99th WS (the only biomechanical metric that exhibited significant association with geometric indices) using Spearman's correlation and subsequently with multivariate linear regression using backward elimination. For the Asian AAA group, 99th WS was highly correlated (R2 = 0.77) with three geometric indices, namely tortuosity, intraluminal thrombus volume, and area-averaged Gaussian curvature. Similarly, 99th WS in the Caucasian AAA group was highly correlated (R2 = 0.87) with six geometric indices, namely maximum AAA diameter, distal neck diameter, diameter-height ratio, minimum wall thickness variance, mode of the wall thickness variance, and area-averaged Gaussian curvature. Significant differences were found between the two groups for ten geometric indices; however, no differences were found for any of their respective biomechanical attributes. Assuming maximum AAA diameter as the most predictive metric for wall stress was found to be imprecise: 24% and 28% accuracy for the Asian and Caucasian groups, respectively. This investigation reveals that geometric indices other than maximum AAA diameter can serve as predictors of wall stress, and potentially for assessment of aneurysm rupture risk, in the Asian and Caucasian AAA populations.

Nonlinear Dose-Response Modeling of High-Throughput Screening Data Using an Evolutionary Algorithm

Nonlinear dose-response relationships exist extensively in the cellular, biochemical, and physiologic processes that are affected by varying levels of biological, chemical, or radiation stress. Modeling such responses is a crucial component of toxicity testing and chemical screening. Traditional model fitting methods such as nonlinear least squares (NLS) are very sensitive to initial parameter values and often had convergence failure. The use of evolutionary algorithms (EAs) has been proposed to address many of the limitations of traditional approaches, but previous methods have been limited in the types of models they can fit. Therefore, we propose the use of an EA for dose-response modeling for a range of potential response model functional forms. This new method can not only fit the most commonly used nonlinear dose-response models (eg, exponential models and 3-, 4-, and 5-parameter logistic models) but also select the best model if no model assumption is made, which is especially useful in the case of high-throughput curve fitting. Compared with NLS, the new method provides stable and robust solutions without sensitivity to initial values.

Nonlinear Regression Operating on Microstructures Described from Topological Data Analysis for the Real-Time Prediction of Effective Properties

Real-time decision making needs evaluating quantities of interest (QoI) in almost real time. When these QoI are related to models based on physics, the use of Model Order Reduction techniques allows speeding-up calculations, enabling fast and accurate evaluations. To accommodate real-time constraints, a valuable route consists of computing parametric solutions—the so-called computational vademecums—that constructed off-line, can be inspected on-line. However, when dealing with shapes and topologies (complex or rich microstructures) their parametric description constitutes a major difficulty. In this paper, we propose using Topological Data Analysis for describing those rich topologies and morphologies in a concise way, and then using the associated topological descriptions for generating accurate supervised classification and nonlinear regression, enabling an almost real-time evaluation of QoI and the associated decision making.

Influence of Sampling Site and Fluid Flow on the Accuracy of Total Body Clearance Calculation

Studies have showed that by assuming arteriovenous drug concentrations are homogenous after intravenous injection, the determination of total body clearance based on venous drug concentrations is often inaccurate. This study considers the use of a fluidic pharmacokinetic profile generator where 28 different profile types were generated corresponding to a physiological model with varying sampling sites, administration locations, and fluid flow rates. Clearance was calculated using established equations, commercial software, and recently proposed models. The results show large differences in clearance values calculated with published equations and commercial software relative to the actual value of clearance. Alterations in sampling site, administration location, and fluid flow rates each influence the extent of calculation errors. The data show that a significant drug concentration gradient exists within the central circulatory system. The results show that the best way to address this issue would be to inject the drug at a peripheral location to allow for sufficient mixing and then sample from a large vein. Extrapolating for missing data can also lead to large errors in clearance calculation; this can be addressed by collecting more samples early after IV bolus administration or by collecting data during steady state conditions for an IV infusion.

Deep Residual Learning for Nonlinear Regression

Deep learning plays a key role in the recent developments of machine learning. This paper develops a deep residual neural network (ResNet) for the regression of nonlinear functions. Convolutional layers and pooling layers are replaced by fully connected layers in the residual block. To evaluate the new regression model, we train and test neural networks with different depths and widths on simulated data, and we find the optimal parameters. We perform multiple numerical tests of the optimal regression model on multiple simulated data, and the results show that the new regression model behaves well on simulated data. Comparisons are also made between the optimal residual regression and other linear as well as nonlinear approximation techniques, such as lasso regression, decision tree, and support vector machine. The optimal residual regression model has better approximation capacity compared to the other models. Finally, the residual regression is applied into the prediction of a relative humidity series in the real world. Our study indicates that the residual regression model is stable and applicable in practice.

The regression Tsetlin machine: a novel approach to interpretable nonlinear regression

Relying simply on bitwise operators, the recently introduced Tsetlin machine (TM) has provided competitive pattern classification accuracy in several benchmarks, including text understanding. In this paper, we introduce the regression Tsetlin machine (RTM), a new class of TMs designed for continuous input and output, targeting nonlinear regression problems. In all brevity, we convert continuous input into a binary representation based on thresholding, and transform the propositional formula formed by the TM into an aggregated continuous output. Our empirical comparison of the RTM with state-of-the-art regression techniques reveals either superior or on par performance on five datasets. This article is part of the theme issue 'Harmonizing energy-autonomous computing and intelligence'.

A design criterion for symmetric model discrimination based on flexible nominal sets

Experimental design applications for discriminating between models have been hampered by the assumption to know beforehand which model is the true one, which is counter to the very aim of the experiment. Previous approaches to alleviate this requirement were either symmetrizations of asymmetric techniques, or Bayesian, minimax, and sequential approaches. Here we present a genuinely symmetric criterion based on a linearized distance between mean‐value surfaces and the newly introduced tool of flexible nominal sets. We demonstrate the computational efficiency of the approach using the proposed criterion and provide a Monte‐Carlo evaluation of its discrimination performance on the basis of the likelihood ratio. An application for a pair of competing models in enzyme kinetics is given.

Evaluation of the paraoxonase-1 kinetic parameters of the lactonase activity by nonlinear fit of progress curves

Although paraoxonase-1 (PON1) activity has been demonstrated to be a reliable biomarker of various diseases, clinical studies have been based only on relative comparison of specific enzyme activities, which capture differences mainly due to (usually unknown) PON1 concentration. Hence, the aim of this report is to present for the first time the simple evaluation method for determining autonomous kinetic parameter of PON1 that could be also associated with polymorphic forms and diseases; i.e. the Michaelis constant which is enzyme concentration independent quantity. This alternative approach significantly reduces the number of experiments needed, and it yields the results with great accuracy.

Accuracy and Precision of the Receptorial Responsiveness Method (RRM) in the Quantification of A1 Adenosine Receptor Agonists

The receptorial responsiveness method (RRM) is a procedure that is based on a simple nonlinear regression while using a model with two variables (X, Y) and (at least) one parameter to be determined (c_x). The model of RRM describes the co-action of two agonists that consume the same response capacity (due to the use of the same postreceptorial signaling in a biological system). While using RRM, uniquely, an acute increase in the concentration of an agonist (near the receptors) can be quantified (as c_x), via evaluating E/c curves that were constructed with the same or another agonist in the same system. As this measurement is sensitive to the implementation of the curve fitting, the goal of the present study was to test RRM by combining different ways and setting options, namely: individual vs. global fitting, ordinary vs. robust fitting, and three weighting options (no weighting vs. weighting by 1/Y² vs. weighting by 1/SD²). During the testing, RRM was used to estimate the known concentrations of stable synthetic A₁ adenosine receptor agonists in isolated, paced guinea pig left atria. The estimates were then compared to the known agonist concentrations (to assess the accuracy of RRM); furthermore, the 95% confidence limits of the best-fit values were also considered (to evaluate the precision of RRM). It was found that, although the global fitting offered the most convenient way to perform RRM, the best estimates were provided by the individual fitting without any weighting, almost irrespective of the fact whether ordinary or robust fitting was chosen.

Gestational growth trajectories derived from a dynamic fetal-placental scaling law

Fetal trajectories characterizing growth rates in utero have relied primarily on goodness of fit rather than mechanistic properties exhibited in utero. Here, we use a validated fetal-placental allometric scaling law and a first principles differential equations model of placental volume growth to generate biologically meaningful fetal-placental growth curves. The growth curves form the foundation for understanding healthy versus at-risk fetal growth and for identifying the timing of key events in utero.

The Use of Correlated Binomial Distribution in Estimating Error Rates for Firearm Evidence Identification

In the branch of forensic science known as firearm evidence identification, estimating error rates is a fundamental challenge. Recently, a new quantitative approach known as the congruent matching cells (CMC) method was developed to improve the accuracy of ballistic identifications and provide a basis for estimating error rates. To estimate error rates, the key is to find an appropriate probability distribution for the relative frequency distribution of observed CMCs overlaid on a relevant measured firearm surface such as the breech face of a cartridge case. Several probability models based on the assumption of independence between cell pair comparisons have been proposed, but the assumption of independence among the cell pair comparisons from the CMC method may not be valid. This article proposes statistical models based on dependent Bernoulli trials, along with corresponding methodology for parameter estimation. To demonstrate the potential improvement from the use of the dependent Bernoulli trial model, the methodology is applied to an actual data set of fired cartridge cases.

Factors Enhancing Serum Syndecan-1 Concentrations: A Large-Scale Comprehensive Medical Examination

Endothelial disorders are related to various diseases. An initial endothelial injury is characterized by endothelial glycocalyx injury. We aimed to evaluate endothelial glycocalyx injury by measuring serum syndecan-1 concentrations in patients during comprehensive medical examinations. A single-center, prospective, observational study was conducted at Asahi University Hospital. The participants enrolled in this study were 1313 patients who underwent comprehensive medical examinations at Asahi University Hospital from January 2018 to June 2018. One patient undergoing hemodialysis was excluded from the study. At enrollment, blood samples were obtained, and study personnel collected demographic and clinical data. No treatments or exposures were conducted except for standard medical examinations and blood sample collection. Laboratory data were obtained by the collection of blood samples at the time of study enrolment. According to nonlinear regression, the concentrations of serum syndecan-1 were significantly related to age (p = 0.016), aspartic aminotransferase concentration (AST, p = 0.020), blood urea nitrogen concentration (BUN, p = 0.013), triglyceride concentration (p < 0.001), and hematocrit (p = 0.006). These relationships were independent associations. Endothelial glycocalyx injury, which is reflected by serum syndecan-1 concentrations, is related to age, hematocrit, AST concentration, BUN concentration, and triglyceride concentration.

Quantifying the Interactions between Biomolecules: Guidelines for Assay Design and Data Analysis

The accurate and precise determination of binding interactions plays a central role in fields such as drug discovery where structure-activity relationships guide the selection and optimization of drug leads. Binding is often assessed by monitoring the response caused by varying one of the binding partners in a functional assay or by using methods where the concentrations of free and/or bound ligand can be directly determined. In addition, there are also many approaches where binding leads to a change in the properties of the binding partner(s) that can be directly quantified such as an alteration in mass or in a spectroscopic signal. The analysis of data resulting from these techniques invariably relies on computer software that enable rapid fitting of the data to nonlinear multiparameter equations. The objective of this Perspective is to serve as a reminder of the basic assumptions that are used in deriving these equations and thus that should be considered during assay design and subsequent data analysis. The result is a set of guidelines for authors considering submitting their work to journals such as ACS Infectious Diseases.

Estimating the Expected Value of Sample Information across Different Sample Sizes Using Moment Matching and Nonlinear Regression

Background. The expected value of sample information (EVSI) determines the economic value of any future study with a specific design aimed at reducing uncertainty about the parameters underlying a health economic model. This has potential as a tool for trial design; the cost and value of different designs could be compared to find the trial with the greatest net benefit. However, despite recent developments, EVSI analysis can be slow, especially when optimizing over a large number of different designs. Methods. This article develops a method to reduce the computation time required to calculate the EVSI across different sample sizes. Our method extends the moment-matching approach to EVSI estimation to optimize over different sample sizes for the underlying trial while retaining a similar computational cost to a single EVSI estimate. This extension calculates the posterior variance of the net monetary benefit across alternative sample sizes and then uses Bayesian nonlinear regression to estimate the EVSI across these sample sizes. Results. A health economic model developed to assess the cost-effectiveness of interventions for chronic pain demonstrates that this EVSI calculation method is fast and accurate for realistic models. This example also highlights how different trial designs can be compared using the EVSI. Conclusion. The proposed estimation method is fast and accurate when calculating the EVSI across different sample sizes. This will allow researchers to realize the potential of using the EVSI to determine an economically optimal trial design for reducing uncertainty in health economic models. Limitations. Our method involves rerunning the health economic model, which can be more computationally expensive than some recent alternatives, especially in complex models.

Comparison of various chemometric analysis for rapid prediction of thiobarbituric acid reactive substances in rainbow trout fillets by hyperspectral imaging technique

This study explores the potential application of hyperspectral imaging (HSI; 430-1,010 nm) coupled with different linear and nonlinear models for rapid nondestructive evaluation of thiobarbituric acid-reactive substances (TBARS) value in rainbow trout (Oncorhynchus mykiss) fillets during 12 days of cold storage (4 ± 2°C). HSI data and TBARS value of fillets were obtained in the laboratory. The primary prediction models were established based on linear partial least squares regression (PLSR) and least squares support vector machine (LS-SVM). In full spectral range, the prediction capability of LS-SVM ( R P 2  = 0.829; RMSEP = 0.128 mg malondialdehyde [MDA]/kg) was better than PLSR ( R P 2  = 0.748; RMSEP = 0.155 mg MDA/kg) model and LS-SVM model exhibited satisfactory prediction performance ( R P 2  > 0.82). To simplify the calibration models, a combination of uninformative variable elimination and backward regression (UB) was used as variable selection. Nine wavelengths were selected. Various chemometric analysis methods including linear PLSR and multiple linear regression and nonlinear LS-SVM and back-propagation artificial neural network (BP-ANN) were compared. The simplified models showed better capability than those were built based on the whole dataset in prediction of TBARS values. Moreover, the nonlinear models were preferred over linear models. Among the four chemometric algorithms, the best and weakest models were LS-SVM and PLSR model, respectively. UB-LS-SVM model was the optimal models for predicting TBARS value in rainbow trout fillets ( R P 2  = 0.831; RMSEP = 0.125 mg MDA/kg). The establishing of lipid-oxidation prediction model in rainbow trout fish was complicated, due to the fluctuations of TBARS values during storage. Therefore, further researches are needed to improve the prediction results and applicability of HIS technique for prediction of TBARS value in rainbow trout fish.

Eddy Current Sensor System for Blade Tip Clearance Measurement Based on a Speed Adjustment Model

Blade tip clearance (BTC) measurement and active clearance control (ACC) are becoming crucial technologies in aero-engine health monitoring so as to improve the efficiency and reliability as well as to ensure timely maintenance. Eddy current sensor (ECS) offers an attractive option for BTC measurement due to its robustness, whereas current approaches have not considered two issues sufficiently. One is that BTC affects the response time of a measurement loop, the other is that ECS signal decays with increasing speed. This paper proposes a speed adjustment model (SAM) to deal with these issues in detail. SAM is trained using a nonlinear regression method from a dynamic training data set obtained by an experiment. The Levenberg–Marquardt (LM) algorithm is used to estimate SAM characteristic parameters. The quantitative relationship between the response time of ECS measurement loop and BTC, as well as the output signal and speed are obtained. A BTC measurement method (BTCMM) based on the SAM is proposed and a geometric constraint equation is constructed to assess the accuracy of BTC measurement. Experiment on a real-time BTC measurement during the running process for a micro turbojet engine is conducted to validate the BTCMM. It is desirable and significative to effectively improve BTC measurement accuracy and expand the range of applicable engine speed.

Physicochemical QSAR Analysis of Passive Permeability Across Caco-2 Monolayers

Caco-2 cell line is frequently used as a simplified in vitro model of intestinal absorption. In this study, a database of 1366 Caco-2 permeability coefficients (P_e) for 768 diverse drugs and drug-like compounds was compiled from public sources. The collected data represent permeation rates measured at varying experimental conditions (pH from 4.0 to 8.0, and stirring rates from 0 to >1000 rpm) that presumably account for passive diffusion across mucosal epithelium. These data were subjected to multistep nonlinear regression analysis using a minimal set of physicochemical descriptors (octanol-water log D, pKa, hydrogen bonding potential, and molecular size). The model was constructed in a mechanistic manner incorporating the following components: (i) a hydrodynamic equation of size- and charge-specific along with nonspecific diffusion across the paracellular pathway; (ii) transcellular diffusion represented by thermodynamic membrane/water partitioning ratio; (iii) stirring-dependent limit of maximum achievable permeability due to the presence of unstirred water layer. The obtained model demonstrates good accuracy of log P_e predictions with a residual mean square error <0.5 log units for all training and validation sets. Given its robust performance and straightforward interpretation in terms of simple physicochemical properties, the proposed model may serve as a valuable tool to guide drug discovery efforts toward readily absorbable compounds.

Dictionary-Free MRI PERK: Parameter Estimation via Regression with Kernels

This paper introduces a fast, general method for dictionary-free parameter estimation in quantitative magnetic resonance imaging (QMRI) parameter estimation via regression with kernels (PERK). PERK first uses prior distributions and the nonlinear MR signal model to simulate many parameter-measurement pairs. Inspired by machine learning, PERK then takes these parameter-measurement pairs as labeled training points and learns from them a nonlinear regression function using kernel functions and convex optimization. PERK admits a simple implementation as per-voxel nonlinear lifting of MRI measurements followed by linear minimum mean-squared error regression. We demonstrate PERK for $ {\textit {T}_{1}}, {\textit {T}_{2}}$ estimation, a well-studied application where it is simple to compare PERK estimates against dictionary-based grid search estimates and iterative optimization estimates. Numerical simulations as well as single-slice phantom and in vivo experiments demonstrate that PERK and other tested methods produce comparable $ {\textit {T}_{1}}, {\textit {T}_{2}}$ estimates in white and gray matter, but PERK is consistently at least $140\times $ faster. This acceleration factor may increase by several orders of magnitude for full-volume QMRI estimation problems involving more latent parameters per voxel.

A two-part mixed effects model for cigarette purchase task data

The Cigarette Purchase Task is a behavioral economic assessment tool designed to measure the relative reinforcing efficacy of cigarette smoking across different prices. An exponential demand equation has become a standard model for analyzing purchase task data, but its utility is compromised by its inability to accommodate values of zero consumption. We propose a two-part mixed effects model that keeps the same exponential demand equation for modeling nonzero consumption values, while providing a logistic regression for the binary outcome of zero versus nonzero consumption. Therefore, the proposed model can accommodate zero consumption values and retain the features of the exponential demand equation at the same time. As a byproduct, the logistic regression component of the proposed model provides a new demand index, the "derived breakpoint", for the price above which a subject is more likely to be abstinent than to be smoking. We apply the proposed model to data collected at baseline from college students (N = 1,217) enrolled in a randomized clinical trial utilizing financial incentives to motivate tobacco cessation. Monte Carlo simulations showed that the proposed model provides better fits than an existing model. We note that the proposed methodology is applicable to other purchase task data, for example, drugs of abuse.

Parametric modeling and optimal experimental designs for estimating isobolograms for drug interactions in toxicology

In toxicology and related areas, interaction effects between two substances are commonly expressed through a combination index [Formula: see text] evaluated separately at different effect levels and mixture ratios. Often, these indices are combined into a graphical representation, the isobologram. Instead of estimating the combination indices at the experimental mixture ratios only, we propose a simple parametric model for estimating the underlying interaction function. We integrate this approach into a joint model where both the parameters of the dose-response functions of the singular substances and the interaction parameters can be estimated simultaneously. As an additional benefit, this concept allows to determine optimal statistical designs for combination studies optimizing the estimation of the interaction function as a whole. From an optimal design perspective, finding the interaction parameters generally corresponds to a [Formula: see text]-optimality resp. [Formula: see text]-optimality design problem, while estimation of all underlying dose response parameters corresponds to a [Formula: see text]-optimality design problem. We show how optimal designs can be obtained in either case as well as how combination designs providing reasonable performance in regard to both criteria can be determined by putting a constraint on the efficiency in regard to one of the criteria and optimizing for the other. As all designs require prior information about model parameter values, which may be unreliable in practice, the effect of misspecifications is investigated as well.