Calibration Set Research Articles

Rapid and non-invasive estimation of total polyphenol content and antioxidant activity of natural corks by NIR spectroscopy and multivariate analysis

The use of near infrared spectroscopy (NIRS) coupled with multivariate data analysis has been evaluated for the rapid and non-invasive estimation of total polyphenol content (TPC) and antioxidant activity of natural corks. A total of 132 samples were used, using two-thirds and one-third of spectra for calibration and external validation sets, respectively. Regression models were developed by partial least squares (PLS) regression analysis. The best models were achieved by different spectral preprocessing approaches with 7 latent variables for TPC and 9 for antioxidant activity, a variance explained by the models between 64.7 % and 72.0 %, and a standard error of calibration (SEC) between 3.6 and 6.8, respectively. Good relationships between chemical values measured by conventional assays and NIRS predicted values were obtained, with standard error of prediction (SEP) of 5.0 and 5.8 for ABTS and DPPH antioxidant activity, and 5.8 for TPC. Regression lines for calibration and external validation presented good correlation coefficients (r2) from 0.80 to 0.85, and from 0.62 to 0.76, respectively. The models obtained showed that it is possible to estimate rapidly and simultaneously, the TPC and antioxidant activity of natural corks by combining the NIRS with different multivariate analysis tools.

Estimating the canopy chlorophyll content of winter wheat under nitrogen deficiency and powdery mildew stress using machine learning

As an important indicator of the photosynthetic capacity of crops, the canopy chlorophyll content (CCC) is nondestructively estimated by reflectance using various spectrometers. Crop growth is often severely affected by Nitrogen (N) deficiency and diseases, and the compatibility of the data collected for different stressors needs further clarification to develop unified estimation models. In this field experimental study, hyperspectral data of the wheat canopy were collected, along with canopy chlorophyll content, to assess nitrogen deficiency and powdery mildew stress. Comparative analysis of hyperspectral remote sensing data input features (original reflectance (OR), spectral index (SI) and wavelet features (WF)) was conducted. A combination of feature selection and machine learning was used to determine the best estimation mode for the accurate inversion of CCC under these two stressors. The results showed that the canopy spectra under nitrogen deficiency and powdery mildew stress had the same change trend. Under nitrogen deficiency, the sensitive wavelengths to CCC mainly reflected canopy structure characteristics, followed by pigments. Under powdery mildew stress, the sensitive wavelengths mainly reflected pigment characteristics, followed by canopy structure. Eight input features (two reflectance wavelengths, two spectral indices and four wavelet features) were selected using competitive adaptive reweighted sampling (CARS) and variance inflation factor (VIF) methods. Machine learning (ML) produced better estimates for both stressors. For CCC estimation under nitrogen stress, random forest regression (RFR) was more suitable (R2 = 0.828; RMSE = 0.363 g/m2) and showed a higher accuracy for both the calibration and validation sets. For CCC estimation under powdery mildew stress, support vector machine regression (SVR) was more suitable (R2 = 0.787; RMSE = 0.126 g/m2), especially when OR and WF data were used as input features. For the unified estimation of CCC under both stressors, WF is most effective as an input feature and good accuracy is achieved for both SVR (R2 = 0.846; RMSE = 0.296 g/m2) and RFR (R2 = 0.844; RMSE = 0.297 g/m2), and their differences are very small. The results demonstrated that the CWT-CARS-VIF-ML mode was appropriate for CCC estimation under two different stressors, which provides an ideal reference and technical guidance for the evaluation of photosynthetic potential and improved crop management.

Quality Assessment and Ripeness Prediction of Table Grapes Using Visible-Near-Infrared Spectroscopy.

Ripeness significantly affects the commercial values and sales of fruits. In order to monitor the change of grapes' quality parameters during ripening, a rapid and nondestructive method of visible-near-infrared spectral (Vis-NIR) technology was utilized in this study. Firstly, the physicochemical properties of grapes at four different ripening stages were explored. Data evidenced increasing color in redness/greenness (a*) and Chroma (C*) and soluble solids (SSC) content and decreasing values in color of lightness (L*), yellowness/blueness (b*) and Hue angle (h*), hardness, and total acid (TA) content as ripening advanced. Based on these results, spectral prediction models for SSC and TA in grapes were established. Effective wavelengths were selected by the competitive adaptive weighting algorithm (CARS), and six common preprocessing methods were applied to pretreat the spectra data. Partial least squares regression (PLSR) was applied to establish models on the basis of effective wavelengths and full spectra. The predictive PLSR models built with full spectra data and 1st derivative preprocessing provided the best values of performance parameters for both SSC and TA. For SSC, the model showed the coefficients of determination for calibration (RCal2) and prediction (RPre2) set of 0.97 and 0.93, respectively, the root mean square error for calibration set (RMSEC) and prediction set (RMSEP) of 0.62 and 1.27, respectively; and the RPD equal to 4.09. As for TA, the optimum values of RCal2, RPre2, RMSEC, RMSEP and RPD were 0.97, 0.94, 0.88, 1.96 and 4.55, respectively. The results indicated that Vis-NIR spectroscopy is an effective tool for the rapid and non-destructive detection of SSC and TA in grapes.

Near infrared spectroscopy for simultaneous quantification of five chemical components in Arnebiae Radix (AR) with partial least squares and support vector machine algorithms

Arnebiae Radix (AR) has high economic and medicinal value, and it is used as Chinese herbal medicine in clinic for a long time. Nowadays, the quality assessment of AR is generally performed on expensive instruments with time-consuming and laborious methods. In this study, the near-infrared (NIR) spectroscopy technique was applied to develop a rapid, simple and accurate method for quantifying five main chemical components including acetylshikonin, β-acetoxyisovalerylalkannin, isobutylshikonin, isovalerylshikonin and β,β-dimethylacrylalkannin in AR coupled with chemometrics, aiming at rapidly evaluating the quality of AR. NIR spectra of 110 batches of AR samples were collected to quantify the content of five chemical components. The partial least squares (PLS) and support vector machine (SVM) algorithms were separately used to develop the calibration models. Different spectral pretreatment methods and feature selection methods were compared in details during the model development. As a result, the performances of PLS models for quantifying the five chemical components were superior to SVM models. The correlation coefficients of calibration set (rcal) of PLS models for acetylshikonin, β-acetoxyisovalerylalkannin, isobutylshikonin, isovalerylshikonin and β,β-dimethylacrylalkannin were 0.9972, 0.9973, 0.9927, 0.9981 and 0.9977, respectively, while the correlation coefficients of prediction set (rpre) were 0.9913, 0.9855, 0.9825, 0.9916 and 0.9825, respectively. Besides, the relative prediction deviation (RPD) of PLS models reached 7.0, 5.85, 4.92, 7.52 and 5.43, respectively. This study demonstrated that the NIR spectroscopy technique was a promising approach as routine analysis of AR, and it could be an alternative method for rapid quality assessment of AR.

Fast and nondestructive discrimination of quality level of Yongchuan Xiuya tea based on near infrared spectroscopy combined with artificial neural network algorithm

Near infrared spectroscopy (NIRS) was used to discriminate the quality level of Yongchuan Xiuya tea quickly and nondestructively. Three quality levels of Yongchuan Xiuya tea were collected, then scanning NIRS, pretreating spectral noise information, screening characteristic spectral intervals by backward interval partial least squares, proceeding principal component analysis. Last, the jump connection nets artificial neural network (J-BP-ANN) with three kinds of transfer functions was applied to establish models. The best pretreated method was the combination of multivariate scattering correction and the first derivative. Six characteristic spectral intervals were screened, which accounting for 27.23% spectral data. The cumulative contribution rate of the first three principal components of the selected characteristic spectra was 97.85%. When the J-BP-ANN calibration set model was established with the tanh function, NIRS model had the best results, whose root mean square error and determination coefficient of the cross validation were 0.953 and 0.031, respectively. The root mean square error and the determination coefficient of the prediction set model were 0.942 and 0.041, respectively. The absolute deviation values of prediction set samples were <0.08. The results showed NIRS can predict the quality levels of Yongchuan Xiuya tea quickly and accurately.

Classical versus chemometrics tools for spectrophotometric determination of fluocinolone acetonide, ciprofloxacin HCl and ciprofloxacin impurity-A in their ternary mixture

Green, simple, accurate and robust univariate and chemometrics assisted UV spectrophotometric approaches have been adopted and validated for concurrent quantification of fluocinolone acetonide (FLU), ciprofloxacin HCl (CIP) together with ciprofloxacin impurity-A (CIP imp-A) in their ternary mixture. Double-divisor ratio spectra derivative (DDRD) method has been used for determination of FLU. On the other hand, the first (D1) and second (D2) derivative approaches have been applied for the quantification of CIP and CIP imp-A, respectively. For the ratio difference (RD), derivative ratio (DR), and mean centering of ratio spectra (MC) methods, CIP and its impurity A have been simultaneously determined. The acquired calibration plots were linear over the concentration range of 0.6–20.0 μg/mL, 1.0–40.0 μg/mL and 1.0–40.0 μg/mL for fluocinolone acetonide, ciprofloxacin HCl, and ciprofloxacin impurity-A, respectively. The chemometrics methods namely; partial least squares (PLS) and artificial neural networks (ANN) were used for the concurrent determination of the three adopted components via using twenty-five mixtures as calibration set and fifteen mixtures as validation one. The investigated approaches were validated in accordance with International Council for Harmonisation (ICH) guidelines, and statistically compared with the official ones. The proposed methods were acceptably applied to the examination of FLU and CIP in their pure powders and pharmaceutical ear drops.

Calibrating Data Mismatches in Deep Learning-Based Quantitative Ultrasound Using Setting Transfer Functions.

Deep learning (DL) can fail when there are data mismatches between training and testing data distributions. Due to its operator-dependent nature, acquisition-related data mismatches, caused by different scanner settings, can occur in ultrasound imaging. As a result, it is crucial to mitigate the effects of these mismatches to enable wider clinical adoption of DL-powered ultrasound imaging and tissue characterization. To address this challenge, we propose an inexpensive and generalizable method that involves collecting a large training set at a single setting and a small calibration set at each scanner setting. Then, the calibration set will be used to calibrate data mismatches by using a signals and systems perspective. We tested the proposed solution to classify two phantoms using an L9-4 array connected to a SonixOne scanner. To investigate generalizability of the proposed solution, we calibrated three types of data mismatches: pulse frequency mismatch, focus mismatch, and output power mismatch. Two well-known convolutional neural networks (CNNs), i.e., ResNet-50 and DenseNet-201, were trained using the ultrasound radio frequency (RF) data. To calibrate the setting mismatches, we calculated the setting transfer functions. The CNNs trained without calibration resulted in mean classification accuracies of around 52%, 84%, and 85% for pulse frequency, focus, and output power mismatches, respectively. By using the setting transfer functions, which allowed a matching of the training and testing domains, we obtained the mean accuracies of 96%, 96%, and 98%, respectively. Therefore, the incorporation of the setting transfer functions between scanner settings can provide an economical means of generalizing a DL model for specific classification tasks where scanner settings are not fixed by the operator.

Determination of corn protein content using near-infrared spectroscopy combined with A-CARS-PLS.

In order to quickly and accurately determine the protein content of corn, a new characteristic wavelength selection algorithm called anchor competitive adaptive reweighted sampling (A-CARS) was proposed in this paper. This method first lets Monte Carlo synergy interval PLS (MC-siPLS) to select the sub-intervals where the characteristic variables exist and then uses CARS to screen the variables further. A-CARS-PLS was compared with 6 methods, including 3 feature variable selection methods (GA-PLS, random frog PLS, and CARS-PLS) and 2 interval partial least squares methods (siPLS and MWPLS). The results showed that A-CARS-PLS was significantly better than other methods with the results: RMSECV=0.0336, R2 c=0.9951 in the calibration set; RMSEP=0.0688, R2 p=0.9820 in the prediction set. Furthermore, A-CARS reduced the original 700-dimensional variable to 23 variables. The results showed that A-CARS-PLS was better than some wavelength selection methods, and it has great application potential in the non-destructive detection of protein content in corn.

‘Specific Normalization’ method for NIRS calibrations

In this study, we present a chemometric procedure called ‘Specific Normalization’. It upgrades the calibration for use in a specific set of new samples by normalizing together the spectra of both the calibration set and a new measured sample set. The proposed procedure is shown for near infra-red spectroscopy measurement of olive oil acidity, as an example. This procedure would need its integration as software application for routine analysis.

GaiaData Release 3

Context.GaiaData Release 3 contains astrometry and photometry results for about 1.8 billion sources based on observations collected by the European Space Agency (ESA)Gaiasatellite during the first 34 months of its operational phase (the same period covered byGaiaearly Data Release 3;GaiaEDR3). Low-resolution spectra for 220 million sources are one of the important new data products included in this release.Aims.In this paper, we focus on the external calibration of low-resolution spectroscopic content, describing the input data, algorithms, data processing, and the validation of the results. Particular attention is given to the quality of the data and to a number of features that users may need to take into account to make the best use of the catalogue.Methods.We calibrated an instrument model to relate meanGaiaspectra to the corresponding spectral energy distributions (SEDs) using an extended set of calibrators: this includes modelling of the instrument dispersion relation, transmission, and line spread functions. Optimisation of the model is achieved through total least-squares regression, accounting for errors inGaiaand external spectra.Results.The resulting instrument model can be used for forward modelling ofGaiaspectra or for inverse modelling of externally calibrated spectra in absolute flux units.Conclusions.The absolute calibration derived in this paper provides an essential ingredient for users ofBP/RPspectra. It allows users to connectBP/RPspectra to absolute fluxes and physical wavelengths.

Estimating flavonoid oxidation potentials: mechanisms and charge-related regression models.

In this paper, I tested our quadratic regression models for the estimation of flavonoid oxidation potentials based on spin populations, the differences in the net atomic charges between a cation and a neutral flavonoid, between a radical and an anion of a flavonoid, and between a radical and a neutral flavonoid on a larger set of flavonoids (N = 35). By including six new flavonoids (5,6,7-trihydroxyflavone, 3,3',4',7-tetrahydroxyflavone, 3,7-dihydroxyflavone, 4',7-dihydroxyflavone, 4',5,7-trihydroxyisoflavone, and 6-hydroxyflavone), we created a respectable calibration set of 35 flavonoids with their oxidation potentials all measured at the same conditions by the same experimentalist. The best model was based on the mean values of the three variables using differences in the net atomic charges (R 2 = 0.970, S.E. = 0.043), which are connected with the three different mechanisms of electrochemical oxidation, SET-PT, SPLET, and HAT.

Lyophilization as an alternative for conservation of equine plasma as a source of IgG for neonatal foals

The administration of plasma to foals with failed transfer of passive immunity is of utmost importance. Transporting refrigerated or frozen plasma to farms can be costly and complicated. Lyophilized plasma can facilitate transport and storage by improving and spreading the technology. Lyophilization is the most commonly used method for preparing dehydrated proteins, which must have adequate stability for an extended storage period at room temperature. This work aimed to evaluate the IgG concentration of fresh equine plasma (EP) and post-lyophilization. Plasma was collected from 6 healthy male horses, age 4±2 years, 356±62kg. An aliquot of EP was lyophilized. The lyophilization process was carried out at the Research Support Center of the Federal Rural University of Pernambuco. Before being lyophilized, samples were frozen at -80°C for 24 h and then processed, in an Alpha 1-4 LD Plus Christ apparatus, a drying chamber temperature of -27°C and vacuum of 0.52 mbar, for 24 hours. The lyophilized EP obtained was stored in well-sealed microtubes and frozen at -4°C for further analysis. The lyophilized EP was resuspended in deionized water to maintain the original volume. The concentration of IgG in fresh EP and lyophilized and resuspended EP was measured. Samples were evaluated using Simple Radial Immunodiffusion kits (IDRS) commercially available forequine serum and colostrum: IDRING® Horse IgG Test, by IDBiotech (Issoire/France). Horse IgG test plates contain an agar gel with a specific antiserum against horse IgG. Plate wells are filled with standards of 200, 100, 50, and 25 μg/mL of horse IgG and the EP and lyophylised EP samples. During diffusion on the agar gel, the antibodies specifically react with the antigen to form a precipitation ring. A linear calibration curve is established to determine the samples’ IgG concentrations from the measured diameters to known concentrations in the standard calibration set. The homogeneity of the variance between treatments was verified by the Shapiro-Wilk test. The results were subjected to analysis of variance (ANOVA) to determine the effects of lyophilization on the IgG concentrations. Significance was declared at P<0.05. When significant differences were tested using the Tukey, significance was set at P<0.05. The IgG concentration in fresh plasma was 8.9±3.2 g/L; after lyophilization, the plasma showed an IgG concentration of 7.1±2.2 g/L. No significant differences were found in the plasma IgG concentrations (P>0.05). The results suggest that lyophilization can be an alternative for conserving EP, since it is a very efficient process to maintain IgG concentration. Further studies should be carried out regarding the administration of plasma in the foal and the transfer of immunity.

Multi wavelength high precision dynamic spectrum detection based on walsh orthogonal coded frequency division multiplexing

In this paper, a Walsh Orthogonal Coded Frequency Division Multiplexing (WFDM) method is proposed for the first time, and a dynamic spectrum (DS) detection system based on this method for non-invasive blood composition analysis is built. The WFDM method can greatly improve the channel utilization and reduce the bandwidth occupation of the DS detection system, making the system realize the collection of multi-wavelength PPG signals with high signal-to-noise ratio (SNR > 75 dB) and the extraction of high-quality DS. Compared with the spectrometer, the SNR of this DS detection system is improved by about 47–62 dB. In addition, this paper used this DS detection system to collect 193 valid patient samples, and established a calibration model between DS data and hemoglobin (HGB) levels in blood. The results shows that the correlation coefficients of the model calibration set and prediction set are 0.9201 and 0.9031, which are close to the clinical requirement.

Assessment of a two-surface plasticity model for hexagonal materials

A computationally efficient two-surface plasticity model is assessed against crystal plasticity. Focus is laid on the mechanical behavior of magnesium alloys in the presence of ductility-limiting defects, such as voids. The two surfaces separately account for slip and twinning such that the constitutive formulation captures the evolving plastic anisotropy and evolving tension-compression asymmetry. For model identification, a procedure is proposed whereby the initial guess is based on a combination of experimental data and computationally intensive polycrystal calculations from the literature. In drawing direct comparisons with crystal plasticity, of which the proposed model constitutes a heuristically derived reduced-order model, the available crystal plasticity simulations are grouped in two datasets. A calibration set contains minimal data for both pristine and porous material subjected to one loading path. Then the two-surface model is assessed against a broader set of crystal plasticity simulations for voided unit cells under various stress states and two loading orientations. The assessment also includes microstructure evolution (rate of growth of porosity and void distortion). The ability of the two-surface model to capture essential features of crystal plasticity is analyzed along with an evaluation of computational cost. The prospects of using the model in guiding the development of physically sound damage models in Mg alloys are put forth in the context of high-throughput simulations.

Improvement approach for determination of cadmium at trace levels in soils by handheld X-ray fluorescence analyzers

Determining trace levels of cadmium in soils by the handheld X-ray fluorescence analyzer is severely hampered by the spectrum baseline. To improve the peak-to-background ratio of the characteristic Cd-Kα X-ray, the filter structure, material, and thickness were optimized; to find the most informative variables for the PLS analysis of trace Cd, the related information enhancement competitive adaptive reweighted sampling (rieCARS) was proposed as a variable selection approach, which contains correlated spectral selection (CSS), double logarithmic processing(DLP), competitive adaptive reweighted sampling (CARS). To eliminate the interference of the K-series line of the Ag anode target on trace Cd, the X-ray tube was used with the Tungsten anode target. One combined filter containing a 0.4 mm thickness Cu and a 0.5 mm thickness Fe was selected, which was placed between the X-ray tube and the sample. The instrumental limit of detection for the Cd descended from 32 mg/kg to 0.02 mg/kg . Sixty-four soil samples were used to establish the model for the filtered signal. The model was compared with PLS, CARS+PLS, DLP + CARS+PLS, CSS + PLS, and CSS + CARS+PLS. The application of rieCARS improved the performance of the calibration model for trace Cd. The coefficient of determination R2 of the rieCARS+PLS model on the calibration set was 0.990; the RMSEP on the test set was 0.28 mg/kg; the average of the relative error in the Poyang Lake soil samples was 0.18; the relative standard deviation for the thirteen replicates of the GSD17 was 0.07. The developed analytical procedure will lay the foundation for the analysis of heavy metal pollution sources in soils and ecological sustainability assessment.

0190 Propensity for Sleep Inertia during Sleep and Its Manifestation After Awakening in a Dynamic Biomathematical Fatigue Model

Abstract Introduction Sleep inertia, the transient period of cognitive impairment experienced immediately after awakening, is of concern in on-call and around-the-clock operational settings and merits development of a predictive tool to manage the associated fatigue risk. Recent work on our previously published biomathematical fatigue model revealed that predictive information about sleep inertia can be gained from tracking the propensity for sleep inertia across time spent asleep. In this approach, a neurobiological process with relatively fast dynamics (in the order of several minutes) interacts with the much slower dynamics of circadian, homeostatic, and allostatic sleep/wake regulation. Here we refine our modeling approach by mathematically dissociating the magnitude of sleep inertia from its dissipation rate. Methods N=280 healthy young adults (ages 21–49y) each participated in one of four laboratory studies of total sleep deprivation, sustained sleep restriction, and/or daytime napping. At 2–4 hour intervals while awake, participants performed a Psychomotor Vigilance Test (PVT), assessing number of lapses (RT&amp;gt;500ms), and rated their sleepiness on the Karolinska Sleepiness Scale (KSS). Sleep periods were recorded polysomnographically. Data were divided into a calibration set used to estimate two new model parameters capturing sleep inertia, and a validation set used to independently verify model goodness-of-fit. Sleep inertia was modeled with two ordinary differential equations – one during wakefulness to track manifest impairment, and one during sleep to track the propensity for sleep inertia upon awakening. Two parameters were fit, one modulating sleep inertia magnitude and one providing a dynamic time constant. Results Based on the calibration data set, the sleep inertia dissipation half-life was found to be 0.43h±0.07h. Based on the validation data set, the goodness-of-fit root-mean-square-error was 2.56 for PVT and 1.05 for KSS, indicating moderate-to-high predictive accuracy. As an emergent property of the model, the propensity of sleep inertia first increased and then decreased over sleep time, peaking 1–2h into the sleep period. Conclusion A two-parameter expansion of our biomathematical fatigue model captured the transient effect of sleep inertia accurately for sleep deprivation, sleep restriction, and daytime napping scenarios, and revealed a novel, emerging dynamic of sleep inertia propensity during sleep. Support (if any) Federal Express Corporation and WSU HPC

An independent validation reveals the potential to predict Hagberg–Perten falling number using spectrometers

AbstractThe Hagberg–Perten falling number (HFN) method is the international standard used to evaluate the damage to wheat (Triticum aestivum) grain quality due to preharvest sprouting (PHS) and late maturity alpha‐amylase (LMA). However, the HFN test requires specialized laboratory facilities and is time consuming. Spectrometers were known as a potential tool for quick HFN assessment, but none of the studies have validated the assessment results across different datasets. In this study, an independent validation was conducted using independent samples and spectral instruments. The calibration set had 462 grain samples of 92 varieties grown at 24 locations in 2019 and examined using a near‐infrared spectrometer. In the validation set, 19 varieties collected from 10 locations in 2 years that experienced either PHS or LMA were scanned with a hyperspectral camera. The association between spectra and HFN was modeled by partial least square regression. As a result, the independent validation correlation accuracy was r = 0.72 and a mean absolute error of 56 s. Furthermore, this study showed a cost‐effective alternative using only 10 spectral bands to predict HFN, and it achieved better performance than the full spectrum of the hyperspectral system. In conclusion, this is the first study that showed the potential that wheat HFN could be predicted on an independent dataset measured by a different instrument. The result suggested that spectrometers can potentially serve as a faster alternative for plant breeders to develop varieties resistant to PHS and LMA, and for growers to screen damaged grains in transportation processes.

Wavelet Analysis and the Information Cost Function Index for Selection of Calibration Events for Flood Simulation

Globally, floods are a prevalent type of natural disaster. Simulating floods is a critical component in the successful implementation of flood management and mitigation strategies within a river basin or catchment area. Selecting appropriate calibration data to establish a reliable hydrological model is of great importance for flood simulation. Usually, hydrologists select the number of flood events used for calibration depending on the catchment size. Currently, there is no numerical index to help hydrologists quantitatively select flood events for calibrating the hydrological models. The question is, what is the necessary and sufficient amount (e.g., 10 events) of calibration flood events that must be selected? This study analyses the spectral characteristics of flood data in Sequences before model calibration. The absolute best set of calibration data is selected using an entropy-like function called the information cost function (ICF), which is calculated from the discrete wavelet transform (DWT) decomposition results. Given that the validation flood events have already been identified, we presume that the greater the similarity between the calibration dataset and the validation dataset, the higher the performance of the hydrological model should be after calibration. The calibration datasets for the Tunxi catchment in southeast China were derived from 21 hourly flood events, and the calibration datasets were generated by arranging 14 flood events in sequences from 3 to 14 (i.e., a Sequence of 3 with 12 sets (set 1 = flood events 1, 2, 3; set 2 = flood events 2, 3, 4, …, and so on)), resulting in a total of 12 sequences and 78 sets. With a predetermined validation set of 7 flood events and the hydrological model chosen as the Hydrologic Engineering Center (HEC–HMS) model, the absolute best calibration flood set was selected. The best set from the Sequence of 10 (set 4 = S10′) was found to be the absolute best calibration set of flood events. The potential of the percentile energy entropy was also analyzed for the best calibration sets, but the ICF was the most consistent index to reveal the ranking based on similarity with model performance. The proposed ICF index in this study is helpful for hydrologists to use data efficiently with more hydrological data obtained in the new era of big data. This study also demonstrates the possibility of improving the effectiveness of utilizing calibration data, particularly in catchments with limited data.

Hyper-parameter tuning and feature extraction for asynchronous action detection from sub-thalamic nucleus local field potentials.

Decoding brain states from subcortical local field potentials (LFPs) indicative of activities such as voluntary movement, tremor, or sleep stages, holds significant potential in treating neurodegenerative disorders and offers new paradigms in brain-computer interface (BCI). Identified states can serve as control signals in coupled human-machine systems, e.g., to regulate deep brain stimulation (DBS) therapy or control prosthetic limbs. However, the behavior, performance, and efficiency of LFP decoders depend on an array of design and calibration settings encapsulated into a single set of hyper-parameters. Although methods exist to tune hyper-parameters automatically, decoders are typically found through exhaustive trial-and-error, manual search, and intuitive experience. This study introduces a Bayesian optimization (BO) approach to hyper-parameter tuning, applicable through feature extraction, channel selection, classification, and stage transition stages of the entire decoding pipeline. The optimization method is compared with five real-time feature extraction methods paired with four classifiers to decode voluntary movement asynchronously based on LFPs recorded with DBS electrodes implanted in the subthalamic nucleus of Parkinson's disease patients. Detection performance, measured as the geometric mean between classifier specificity and sensitivity, is automatically optimized. BO demonstrates improved decoding performance from initial parameter setting across all methods. The best decoders achieve a maximum performance of 0.74 ± 0.06 (mean ± SD across all participants) sensitivity-specificity geometric mean. In addition, parameter relevance is determined using the BO surrogate models. Hyper-parameters tend to be sub-optimally fixed across different users rather than individually adjusted or even specifically set for a decoding task. The relevance of each parameter to the optimization problem and comparisons between algorithms can also be difficult to track with the evolution of the decoding problem. We believe that the proposed decoding pipeline and BO approach is a promising solution to such challenges surrounding hyper-parameter tuning and that the study's findings can inform future design iterations of neural decoders for adaptive DBS and BCI.

Cheese yield and nutrients recovery in the curd predicted by Fourier-transform spectra from individual sheep milk samples

The objectives of this study were to explore the use of Fourier-transform infrared (FTIR) spectroscopy on individual sheep milk samples for predicting cheese-making traits, and to test the effect of the farm variability on their prediction accuracy. For each of 121 ewes from 4 farms, a laboratory model cheese was produced, and 3 actual cheese yield traits (fresh cheese, cheese solids, and cheese water) and 4 milk nutrient recovery traits (fat, protein, total solids, and energy) in the curd were measured. Calibration equations were developed using a Bayesian approach with 2 different scenarios: (1) a random cross-validation (80% calibration; 20% validation set), and (2) a leave-one-out validation (3 farms used as calibration, and the remaining one as validation set) to assess the accuracy of prediction of samples from external farms, not included in calibration set. The best performance was obtained for predicting the yield and recovery of total solids, justifying for the practical application of the method at sheep population and dairy industry levels. Performances for the remaining traits were lower, but still useful for the monitoring of the milk processing in the case of fresh curd and recovery of energy. Insufficient accuracies were found for the recovery of protein and fat, highlighting the complex nature of the relationships among the milk nutrients and their recovery in the curd. The leave-one-out validation procedure, as expected, showed lower prediction accuracies, as a result of the characteristics of the farming systems, which were different between calibration and validation sets. In this regard, the inclusion of information related to the farm could help to improve the prediction accuracy of these traits. Overall, a large contribution to the prediction of the cheese-making traits came from the areas known as "water" and "fingerprint" regions. These findings suggest that, according to the traits studied, the inclusion of water regions for the development of the prediction equation models is fundamental to maintain a high prediction accuracy. However, further studies are necessary to better understand the role of specific absorbance peaks and their contribution to the prediction of cheese-making traits, to offer reliable tools applicable along the dairy ovine chain.