Lower Prediction Error Research Articles

Computational prediction of pharmacokinetic parameters as an in vitro approach for assessing paracetamol tablets for IVIVC; a strategy in COVID-19 disruptive times.

In vitro-in vivo correlation (IVIVC) is a desirable attribute for any drug dissolution test to establish relevance and confidence in evaluating the quality and safety of products. The pharmacokinetic parameters of paracetamol have been studied extensively but information about the IVIVC is scanty and mostly controversial; this justifies its choice as a model drug for this study. This work is aimed to evaluate and compare the IVIVC dissolution profile of different brands of paracetamol tablets using authentic pharmacokinetic parameters such as; maximum observed dug concentration (Cmax), time to reach Cmax (Tmax) and Area under Curve concentration-time curve (AUC) obtained from literature. In vitro release data were obtained for each brand (n=12) using the USP II apparatus at 50 rpm in 900 ml phosphate buffer of pH 5.8, maintained at 37±0.5 °C and the results were mathematically extrapolated to predict in vivo data. The percent predicted error (% PE) for Cmax ranges from 1.70 to 6.52 % across the brands, while those for Tmax and AUC were < 0 % and > 20 % respectively. The observed low prediction error for Cmax and Tmax (<10 %) demonstrated that the paracetamol IVIVC model was valid based on FDA guidelines. While a satisfactory result could not be achieved for AUC, promising results were obtained exploiting the convolution based IVIVC model.

A Prediction of Fingertip Force and Joint Kinematics of Individual Fingers using Motoneuron Firing Activities

<h3>Research Objectives</h3> Advanced assistive robotic hands can help individuals with hand impairment to perform daily activities. The detection of intended motor output of individual fingers, however, remains a challenge. The objective of this study was to develop a decoding method to continuously and simultaneously predict fingertip forces and joint kinematics of individual fingers using motoneuron firing frequency. <h3>Design</h3> Blind source separation method was applied to EMG signals to obtain motor unit (MU) firing activities. MU refinement procedure was used to determine finger- and task-specific MU pools. Obtained MU pools were tested on different datasets involving both isometric and dynamic tasks (combined). Populational MU firing frequencies were calculated to estimate the fingertip forces and joint angles concurrently using regression models. Model performance was compared with conventional EMG amplitude-based approach. <h3>Setting</h3> The study was carried out at the NC State/UNC Joint Department of Biomedical Engineering. <h3>Participants</h3> Seven healthy volunteers (Age: 28±7) participated in this study. No participant had any prior neuro-muscular ailments. All participants provided an informed consent. <h3>Interventions</h3> Not applicable. <h3>Main Outcome Measure(s)</h3> Concurrent and continuous prediction of finger kinetics and kinematics via motoneuron activities. <h3>Results</h3> MU approach led to better performance (higher correlation between the predicted and measured output and lower prediction error) in predicting both joint angles and fingertip forces than the EMG amplitude approach. <h3>Conclusion(s)</h3> MU pools can be identified for specific finger-task combinations, which allows concurrent control of fingertip forces and joint angles with assistive robot hands. Separation matrix derived from single finger trials was used applied to combined trials to avoid the computationally intensive nature of HD-EMG decomposition and to determine its usability in real-time. Further development of this method could provide concurrent and continuous control of individual finger movements of highly dexterous robotic hands.

Decentralized Wind Power Forecasting Method Based on Informer

Background: Improving the accuracy of decentralized wind power forecasting has become an important means to ensure the safe and stable operation of the power grid. Traditional prediction methods cannot meet the requirements of precise wind power prediction for long time series Methods: A decentralized wind power prediction model based on Informer is established. Introducing the model's sparse self-attention mechanism, self-attention distillation mechanism and The generative decoder model realizes the long-term sequence power prediction of decentralized wind power. Results: The results of the proposed model are analyzed by comparing it with traditional neural network models such as LSTM and feed-forward neural networks. The root mean square error (RMSE) can be reduced by up to 6.3% Conclusion: It is proved that the Informer model prediction has a lower prediction error and performs better in long-term sequence power pre-tasks.

ML algorithms to estimate data reliability metric of ECG from inter-patient data for trustable AI-based cardiac monitors

Improvement in remote healthcare systems has been conducted utilizing cyber-physical systems (CPS), although data reliability is still the biggest challenge in unsupervised settings. Cardiac monitoring remotely using wearable sensors and implementing trustable artificial intelligence (AI) algorithms are one of the most challenging problems in Smart Health (sHealth). Remote cardiac monitoring using wearable electrocardiogram (ECG) devices is prone to noise and artifacts, which can make the data unreliable as false detection, can happen due to the presence of noise. Checking the reliability of data manually is time-consuming and not possible for long-term recorded data from unsupervised settings. To mitigate the data reliability issue in remote cardiac monitoring, we propose a novel Data Reliability Metric (DReM), where we predict the reliability of inter-patient ECG data using machine learning (ML) algorithms from data statistics itself. We predicted DReM on a scale of 0–1, where 0 means an extremely noisy signal and 1 means a noise-free signal. In this work, we have used Lasso regression (LR), Support Vector Regression (SVR), Decision Tree Regression (DTR), and Random Forest Regression (RFR) algorithms. To evaluate the model performance R 2 , Mean Absolute Error (MAE), and Root Mean Square Error (RMSE) was used. Decision tree regression was able to predict the DReM with a high correlation (R 2 = 1) and low prediction error. We tested our model using different patients’ data and observed that DTR performed best in all cases. Our proposed data reliability approach can be beneficial in cardiac disease detection by identifying false predictions from unreliable data. • A novel Data Reliability Metric (DReM) is proposed to estimate the reliability of inter-patient ECG data on a scale of 0–1. • Both time-series features (statistical, spectral, and temporal) and morphological features were extracted from the noise-contaminated ECG data. • For evaluating model performance R 2 , RMSE, and MAE are used as analysis indices. • Decision tree regression (DTR) is the best-performed model that achieved R 2 value 1. • Our proposed automated method can be useful in remote cardiac monitoring to identify false alarms due to the less reliable data in ECG.

Short-Term Visibility Prediction Using Tree-Based Machine Learning Algorithms and Numerical Weather Prediction Data

Abstract Accurate visibility prediction is imperative in the interests of human and environmental health. However, the existing numerical models for visibility prediction are characterized by low prediction accuracy and high computational cost. Thus, in this study, we predicted visibility using tree-based machine learning algorithms and numerical weather prediction data determined by the local data assimilation and prediction system (LDAPS) of the Korea Meteorological Administration. We then evaluated the accuracy of visibility prediction for Seoul, South Korea, through a comparative analysis using observed visibility from the automated synoptic observing system. The visibility predicted by machine learning algorithm was compared with the visibility predicted by LDAPS. The LDAPS data employed to construct the visibility prediction model were divided into learning, validation, and test sets. The optimal machine learning algorithm for visibility prediction was determined using the learning and validation sets. In this study, the extreme gradient boosting (XGB) algorithm showed the highest accuracy for visibility prediction. Comparative results using the test sets revealed lower prediction error and higher correlation coefficient for visibility predicted by the XGB algorithm (bias: −0.62 km, MAE: 2.04 km, RMSE: 2.94 km, and R: 0.88) than for that predicted by LDAPS (bias: −0.32 km, MAE: 4.66 km, RMSE: 6.48 km, and R: 0.40). Moreover, the mean equitable threat score (ETS) also indicated higher prediction accuracy for visibility predicted by the XGB algorithm (ETS: 0.5–0.6 for visibility ranges) than for that predicted by LDAPS (ETS: 0.1–0.2).

Approaches for the short-term prediction of natural daily streamflows using hybrid machine learning enhanced with grey wolf optimization

ABSTRACT This paper presents the development of hybrid machine learning models to forecast the natural flows of water bodies. Five models were considered under the analysis: extreme gradient boosting (XGB), extreme learning machine (ELM), support vector regression (SVR), elastic net linear model (EN), and multivariate adaptive regression splines (MARS). The grey wolf optimization algorithm (GWO) optimized all of the models’ internal parameters. A feature selection approach was embedded in the hybrid model to reduce the number of input variables. The hybrid model performed the forecasts considering one, three, five, and seven days ahead on data collected from Cahora Bassa dam, Mozambique. In the experiments conducted in this paper, XGB outperformed EN, ELM, MARS, and SVR, presenting lower prediction error and uncertainty. The proposed XGB model arises as an alternative to help with flow prediction, which is crucial for hydroelectric power plant activity.

Synthesizing Images of Tau Pathology from Images of Glucose Utilization

AbstractBackgroundIn vivo tau‐positron emission tomography (PET) is an attractive biomarker for Alzheimer’s disease (AD) diagnosis and treatment. However, tau‐PET is less widely available than other modalities. In this study, we tested cross‐modality synthesis of tau‐PET brain images from fluorodeoxyglucose F‐18 (FDG)‐PET using a deep convolutional neural network (CNN).MethodParticipants (n=1,192) who had brain FDG‐PET with 18F‐FDG and tau‐PET with Flortaucipir (F‐18‐AV‐1451) were included for training and testing. This cohort spanned normal aging (ages 26‐98), pre‐clinical, and clinical AD and related disorders including the FTD and DLB spectrum. External validation was done using ADNI (n=288). The PET scans were co‐registered to the corresponding MRI and subsequently warped to Mayo Clinic Adult Lifespan Template (MCALT) space. Tau‐PET images were SUVR‐normalized to the cerebellar crus, and FDG to the pons. A 3D dense‐U‐net model was utilized as an architecture. Cross‐validation experiments were conducted using 5‐fold validations (60% training set, 20% validation set, and 20% test set) with mean squared error as the loss function.ResultOur dense‐U‐net model successfully synthesized tau‐PET from metabolic images with good correlation and low prediction error for regional SUVRs (Figure 1A‐C). The model showed a robust prediction ability, performing accurately in an independent, external ADNI cohort (Figure 1D‐F). The model‐imputed tau‐PET significantly improved performance in classifying tau positivity (mean AUROC(±SD)=0.78±0.04 and 0.85±0.03 for FDG‐PET and synthesized tau‐PET, respectively) and diagnostic groups (cognitively unimpaired with abnormal amyloid‐PET vs. cognitively impaired with abnormal amyloid‐PET) compared to the original input FDG data (mean AUROC(±SD)=0.89±0.04, 0.85±0.05 0.91±0.04 for actual tau‐PET, FDG‐PET and synthesized tau‐PET, respectively) (Figure 2), suggesting enhanced clinical utility for metabolic images. The ADNI cohort also showed similar results (for tau positivity: AUROC=0.66 and 0.78 for FDG‐PET and synthesized tau‐PET, respectively; for CU A+ vs. CI A+: AUROC=0.86, 0.62, and 0.73 for actual tau‐PET, FDG‐PET, and synthesized tau‐PET; Figure 3).ConclusionWe showed that using a CNN model to predict tau‐PET from FDG‐PET is feasible. The synthesized tau‐PET can augment the value of FDG‐PET, facilitating the multi‐modal diagnosis of AD.

Efficacy Evaluation of Statistical Prediction Models: A Comparative Analysis based on ARIMA Model, Grey Model and Polynomial Regression Model

This paper uses ARIMA model, grey model and polynomial regression model to estimate and forecast two important indicators of China's real GDP growth rate and consumer price index (CPI). The results show that the sequence predicted by the polynomial regression model has the highest degree of agreement with the actual value, and has the lowest prediction error and the best prediction performance, while the other two types of models are not suitable for long-term prediction.

County-scale crop yield prediction by integrating crop simulation with machine learning models.

Crop yield prediction is of great importance for decision making, yet it remains an ongoing scientific challenge. Interactions among different genetic, environmental, and management factors and uncertainty in input values are making crop yield prediction complex. Building upon a previous work in which we coupled crop modeling with machine learning (ML) models to predict maize yields for three US Corn Belt states, here, we expand the concept to the entire US Corn Belt (12 states). More specifically, we built five new ML models and their ensemble models, considering the scenarios with and without crop modeling variables. Additional input values in our models are soil, weather, management, and historical yield data. A unique aspect of our work is the spatial analysis to investigate causes for low or high model prediction errors. Our results indicated that the prediction accuracy increases by coupling crop modeling with machine learning. The ensemble model overperformed the individual ML models, having a relative root mean square error (RRMSE) of about 9% for the test years (2018, 2019, and 2020), which is comparable to previous studies. In addition, analysis of the sources of error revealed that counties and crop reporting districts with low cropland ratios have high RRMSE. Furthermore, we found that soil input data and extreme weather events were responsible for high errors in some regions. The proposed models can be deployed for large-scale prediction at the county level and, contingent upon data availability, can be utilized for field level prediction.

Surrogate Model Development for Optimized Blended-Wing-Body Aerodynamics

In the conceptual design phase of conventional-configuration aircraft, calibrated low-fidelity methods provide sufficiently accurate estimates of aerodynamic coefficients. It has been observed, however, that for blended-wing–body aircraft, important flow effects are not captured adequately with low-fidelity aerodynamic tools. Consequently, high-fidelity methods become necessary to study blended-wing–body aerodynamics. Since repeated function calls are needed in an optimization loop, high-fidelity analysis is prohibitively expensive in the conceptual design phase, where several optimization scenarios are considered. In this paper, the integration of high-fidelity data for blended-wing–body aircraft for a mission calculation module is presented. A surrogate model based on Gaussian processes (GPs) with acceptably low prediction error is sought as an alternative to Reynolds-averaged Navier–Stokes computational fluid dynamics. Three adaptations are considered: sparse GPs, mixtures of GP approximators, and need-based filtering for GP. The results provide benchmark values for this case and show that the combination of subsonic and transonic behaviors in the training set is problematic and that, for the considered datasets, sparse GP models suffer from oversmoothing, while mixtures of GP models suffer from overfitting. From the error levels, it is observed that a GP with an infinitely differentiable squared exponential kernel based on reduced data pertinent to mission analysis is the most effective option.

RSSI Fingerprint Height Based Empirical Model Prediction for Smart Indoor Localization.

Smart indoor living advances in the recent decade, such as home indoor localization and positioning, has seen a significant need for low-cost localization systems based on freely available resources such as Received Signal Strength Indicator by the dense deployment of Wireless Local Area Networks (WLAN). The off-the-shelf user equipment (UE's) available at an affordable price across the globe are well equipped with the functionality to scan the radio access network for hearable single strength; in complex indoor environments, multiple signals can be received at a particular reference point with no consideration of the height of the transmitter and possible broadcasting coverage. Most effective fingerprinting algorithm solutions require specialized labor, are time-consuming to carry out site surveys, training of the data, big data analysis, and in most cases, additional hardware requirements relatively increase energy consumption and cost, not forgetting that in case of changes in the indoor environment will highly affect the fingerprint due to interferences. This paper experimentally evaluates and proposes a novel technique for Received Signal Indicator (RSSI) distance prediction, leveraging transceiver height, and Fresnel ranging in a complex indoor environment to better suit the path loss of RSSI at a particular Reference Point (RP) and time, which further contributes greatly to indoor localization. The experimentation in different complex indoor environments of the corridor and office lab during work hours to ascertain real-life and time feasibility shows that the technique's accuracy is greatly improved in the office room and the corridor, achieving lower average prediction errors at low-cost than the comparison prediction algorithms. Compared with the conventional prediction techniques, for example, with Access Point 1 (AP1), the proposed Height Dependence Path-Loss (HEM) model at 0 dBm error attains a confidence probability of 10.98%, higher than the 2.65% for the distance dependence of Path-Loss New Empirical Model (NEM), 4.2% for the Multi-Wall dependence on Path-Loss (MWM) model, and 0% for the Conventional one-slope Path-Loss (OSM) model, respectively. Online localization, amongst the hearable APs, it is seen the proposed HEM fingerprint localization based on the proposed HEM prediction model attains a confidence probability of 31% at 3 m, 55% at 6 m, 78% at 9 m, outperforming the NEM with 26%, 43%, 62%, 62%, the MWM with 23%, 43%, 66%, respectively. The robustness of the HEM fingerprint using diverse predicted test samples by the NEM and MWM models indicates better localization of 13% than comparison fingerprints.

Investigation on C and ESR Estimation of DC-Link Capacitor in Maglev Choppers Using Artificial Neural Network

The reliability of capacitors is one of the most important issues in power electronics. The health status of capacitors can be evaluated through the comparison of estimated C/ESR values with their original values. In this paper, a two-input artificial neural network (ANN) is proposed for C and ESR estimation of DC-link capacitors in Maglev choppers; combined with the existing voltage and current sensors that are used for protection and control, it provides a promising solution for the health condition monitoring of the Maglev chopper in the Maglev train. Compared with prior-art research, the actual capacitor degradation progress where both C and ESR degrade is considered in training. Moreover, ANN’s advantage of fitting nonlinear and complex relationships is explored by building an aggressive mapping between the voltage ripple at 5 kHz to C/ESR at 120 Hz, which cannot be described or analyzed by linear circuit equations. Thus, cross validation must be implemented to avoid an occasional poor fitting, and ensure the stability of ANN. Experimental results show that the proposed ANN outperforms the support vector regression in both ESR and C estimation. While C estimation suffered from over-fitting and instability, ESR estimation by the ANN is accurate and stable with a low average prediction error within 1%, showing great potential for condition monitoring of DC-link capacitors in Maglev choppers.

A Prediction Method with Data Leakage Suppression for Time Series

In view of the characteristics of the collected time series, such as being high noise, non-stationary and nonlinear, most of the current methods are designed to smooth or denoise the whole time series at one time and then divide the training set and testing set, which will lead to using the information of the testing set in the training process, resulting in data leakage and other problems. In order to reduce the impact of noise on time series prediction and prevent data leakage, a prediction method with data leakage suppression for time series (DLS) is proposed. This prediction method carries out multiple variational mode decomposition on the time series by overlapping slicing and improves the noise reduction threshold function to perform noise reduction processing on the decomposed time series. Furthermore, the idea of deep learning is introduced to establish a neural network multi-step prediction model, so as to improve the performance of time series prediction. The different datasets are selected as experimental data, and the results show that the proposed method has a better prediction effect and lower prediction error, compared with the common multi-step prediction methods, which verifies the superiority of the prediction method.

Optimization of Ultrasound-Assisted Extraction from Young Coconut Mesocarp in the Rapid Extraction of Phenolic Compounds and Antioxidant Activity

Coconut (Cocos nucifera) fruit has attracted consumer interest due to its health benefits, especially the consumption of coconut water produced from 6–7-month-old coconut fruit. Consequently, by-products from young coconut fruit are also being expanded, including coconut mesocarp containing phenolic compounds. Therefore, this study aimed to provide new applications for young coconut mesocarp wastes. Specifically, optimizing the ultrasound-assisted extraction (UAE) of phenolic compounds from coconut mesocarp using a Box–Behnken design in conjunction with response surface methodology (RSM). The effects of three extraction variables, such as temperature (10–70 °C), solvent-to-sample ratio (20:1–10:1 mL g−1), and pulse duty cycle (0.4–1.0 s−1) were examined on the level of total phenolic content (TPC) and antioxidant activity of the extract. Subsequently, the optimum UAE condition was predicted using RSM models with coefficients of determination (R2) higher than 0.94, low prediction errors (less than 2.34), and non-significant lack-of-fit values (p &lt; 0.05) for the two responses. Extraction time was evaluated through kinetic (5 to 25 min) studies applying the optimum extraction temperature (70 °C), solvent-to-sample ratio (20:1 mL g−1), and pulse duty cycle (0.55 s−1). An efficient extraction was achieved within 5 min, resulting in an extract with 47.78 ± 1.24 mg GAE 100 g−1 DW for the total phenolic compounds and high antioxidant activity (87.28 ± 1.01% DPPH). Extraction by ultrasound was then concluded to facilitate a fast extraction rate with high reproducibility (coefficients of variation were less than 3% in the levels of antioxidant activity and phenolic compounds).

Detection and quantification using ATR-FTIR spectroscopy of whey protein concentrate adulteration with wheat flour

Expensive and time-consuming methods are currently used to identify manipulated whey protein concentrate (WPC). We tested the application of attenuated total reflection Fourier-transform infrared (ATR-FTIR) spectroscopy to detect and quantify adulteration of WPC with wheat flour (WF). ATR-FTIR is based on the interaction between IR radiation and the molecular bonds in the samples, allowing the capture of its vibration energy and the acquisition of spectra with information that can be used to identify and quantify functional groups. WPC Samples were adulterated with different percentages of WF and spectra were collected. Amide I and II bands exhibit a decreasing protein content when more WF is added. Conversely, at the carbohydrate characteristic band (1080 cm−1), there is an increasing intensity as more WF is added. Partial least squares (PLS) regression models presented low prediction errors and high coefficients of determination. Furthermore, one of the models was chosen for a blind prediction test of 10 samples adulterated with random amounts of WF, and it could make reasonable predictions of the actual levels of adulteration. Therefore, ATR-FTIR spectroscopy coupled with multivariate analysis shows strong potential to detect adulteration in WPC with WF and the capability to quantify the added mass content.

Predicting ammonia nitrogen in surface water by a new attention-based deep learning hybrid model

Ammonia nitrogen (NH3–N) is closely related to the occurrence of cyanobacterial blooms and destruction of surface water ecosystems, and thus it is of great significance to develop predictive models for NH3–N. However, traditional models cannot fully consider the complex nonlinear relationship between NH3–N and various relative environmental parameters. The long short-term memory (LSTM) neural network can overcome this limitation. A new hybrid model BC-MODWT-DA-LSTM was proposed based on LSTM combining with the dual-stage attention (DA) mechanism and boundary corrected maximal overlap discrete wavelet transform (BC-MODWT) data decomposition method. By introducing attention mechanism, LSTM could selectively focus on the input data. BC-MODWT could decompose the input data into sublayers to determine the main swings and trends of the input feature series. The BC-MODWT-DA-LSTM hybrid model was superior to other studied models with lower average prediction errors. It could maintain NASH Sutcliffe efficiency coefficient (NSE) values above 0.900 under the lead time up to 7 days, and the area under the receiver operating characteristic (ROC) curve could reach 0.992. The hybrid model also had higher prediction accuracies at the peak spots, indicating that it was capable of early warning when sudden high NH3–N pollution occurred. The high forecasting accuracy of the suggested hybrid method proved that further improving LSTM model without introducing more complex topologies was a promising water quality prediction method.

Integration of NIR spectroscopy and chemometrics for authentication and quantitation of adulteration in sweet marjoram (Origanum majorana L.)

Origanum majorana L. (marjoram) has considerable promise as a folk remedy for digestive ailments in addition to its ethnomedicinal use in gynecological disorders. Due to the high demand for the botanical raw material, it is prone to intentional and inadvertent adulteration by morphologically similar herbs. In the current study, the adulteration of marjoram with its common adulterants; oregano, thyme, olive, basil and sage has been thoroughly investigated. To distinguish between marjoram and its prevalent adulterants, an integrated near infrared (NIR) spectroscopic method coupled with chemometric data analysis was developed. For exploratory pattern recognition, unsupervised multivariate models such as principal component analysis (PCA) and hierarchical cluster analysis (HCA) were implemented, followed by supervised models. Authentication of marjoram was successfully assessed using soft independent modelling of class analogy (SIMCA) with 100% sensitivity. Orthogonal projection to latent structures-discriminant analysis (OPLS-DA) modeling was employed to discriminate pure marjoram from its deliberately adulterated mixtures. Further, amounts of the adulterants were quantified in the botanical mixtures using partial least squares regression (PLS-R). The calibration and validation goodness of fit were determined to be greater than 0.9 and almost one, with a low root mean square error of prediction (RMSEP) and high ratio of performance to deviation (RPD) and range error ratio (RER) which highlight the model significant potential. Moreover, the optimal number of latent variables was found to be 5 as computed by permutation test. In the current study, the validity and reliability of the models employed to assess marjoram authenticity and purity was evidenced by both internal and external validation methods with a little to no data pre-processing.

A smartphone-based innovative approach to discriminate red pigments in roman frescoes mock-ups

• Frescoes replicas using red pigments were prepared to develop a smartphone-based methodological approach. • Colour properties were measured by spectrocolorimeter and smartphone. • Chemometric classificatory methods were used to identify red pigments colors. • Low classification errors were obtained employing LDA and SVC classifying methods. • A new and easy approach employing mobile phones to discriminate red pigments is presented. The characterization of red pigments in frescoes wall paintings has been of great interest for researchers to better understand raw material procurement dynamics, pigment receipts, stylistic evolution and to assess their conservation state. In this study a non-destructive colorimetric approach implementing a smartphone-based method was developed in order to be able to distinguish between three pigments made from minium, haematite and cinnabar minerals, and also mixed pigments, preparing frescoes mock-ups following the roman receipt described by Vitruvius. Portable FT-IR, Raman spectroscopy, portable XRF and visible reflectance spectra analyses were carried out as reference methods for smartphone colorimetry results validation. Employing a reference colour sheet to control changing lighting conditions, different chemometric approaches have been developed and tested, cross-referencing standard analytical results with the data obtained by smartphone. Overall, using only colour parameters from the smartphone, a Linear Discriminant Analysis and a Support Vector Classifier were tested to efficiently classify each sample based on the red pigment used, with low prediction errors. This work shows the potential of smartphones as cheap, fast and user-friendly analytical devices for the screening of frescoes, and as a prior selective step before carrying out further more expensive and specialized analyses.

Principal Component Pursuit for Pattern Identification in Environmental Mixtures.

Environmental health researchers often aim to identify sources or behaviors that give rise to potentially harmful environmental exposures. We adapted principal component pursuit (PCP)-a robust and well-established technique for dimensionality reduction in computer vision and signal processing-to identify patterns in environmental mixtures. PCP decomposes the exposure mixture into a low-rank matrix containing consistent patterns of exposure across pollutants and a sparse matrix isolating unique or extreme exposure events. We adapted PCP to accommodate nonnegative data, missing data, and values below a given limit of detection (LOD). We simulated data to represent environmental mixtures of two sizes with increasing proportions and three noise structures. We applied PCP-LOD to evaluate its performance in comparison with principal component analysis (PCA). We next applied principal component pursuit with limit of detection (PCP-LOD) to an exposure mixture of 21 persistent organic pollutants (POPs) measured in 1,000 U.S. adults from the 2001-2002 National Health and Nutrition Examination Survey (NHANES). We applied singular value decomposition to the estimated low-rank matrix to characterize the patterns. PCP-LOD recovered the true number of patterns through cross-validation for all simulations; based on an a priori specified criterion, PCA recovered the true number of patterns in 32% of simulations. PCP-LOD achieved lower relative predictive error than PCA for all simulated data sets with up to 50% of the data . When 75% of values were , PCP-LOD outperformed PCA only when noise was low. In the POP mixture, PCP-LOD identified a rank-three underlying structure and separated 6% of values as extreme events. One pattern represented comprehensive exposure to all POPs. The other patterns grouped chemicals based on known structure and toxicity. PCP-LOD serves as a useful tool to express multidimensional exposures as consistent patterns that, if found to be related to adverse health, are amenable to targeted public health messaging. https://doi.org/10.1289/EHP10479.

A hybrid feature selection algorithm using simplified swarm optimization for body fat prediction

Background and objectivesObesity is one of the chronic diseases that seriously threaten people's health outcomes globally. Since the prevalence of obesity is increasing among people of all ages, measuring the body fat percentages is vital before treatment. However, the body fat percentage cannot be accurately measured by weighing. While many devices are commonly used to measure the body fat percentage, these devices are expensive and depend on complex instruments. Therefore, more practical and cost-effective solutions are desired to measure body fat accurately. This study presents a hybrid feature selection method based on a VIKOR-based multi-filter ensemble technique (VMFET) and an improved simplified swarm optimization (iSSO) to predict the body fat percentage with low prediction error. MethodsThe study followed a two-phase process. First, VMFET was used to aggregate the statistical outcomesof individual filters to filter the most informative features from the original dataset. Then, the selected features are applied to the next phase. Second, iSSO was tailored with a biased random initialization scheme, effect-based feature pruning scheme, and multiple linear regression as a wrapper method to improve the prediction performance and select the optimal feature subset. Studies resultsExtensive experiments were performed using nine datasets to verify the performance of the proposed method empirically, and the corresponding results were compared with up-to-date studies. ConclusionThe statistical results demonstrated that the proposed method offers a promising and effective tool for predicting body fat. SignificanceThe hybrid feature selection model can enhance prediction accuracy and lower prediction error.