RESUMO
This study proposed a noninvasive blood glucose estimation system based on dual-wavelength photoplethysmography (PPG) and bioelectrical impedance measuring technology that can avoid the discomfort created by conventional invasive blood glucose measurement methods while accurately estimating blood glucose. The measured PPG signals are converted into mean, variance, skewness, kurtosis, standard deviation, and information entropy. The data obtained by bioelectrical impedance measuring consist of the real part, imaginary part, phase, and amplitude size of 11 types of frequencies, which are converted into features through principal component analyses. After combining the input of seven physiological features, the blood glucose value is finally obtained as the input of the back-propagation neural network (BPNN). To confirm the robustness of the system operation, this study collected data from 40 volunteers and established a database. From the experimental results, the system has a mean squared error of 40.736, a root mean squared error of 6.3824, a mean absolute error of 5.0896, a mean absolute relative difference of 4.4321%, and a coefficient of determination (R Squared, R2) of 0.997, all of which fall within the clinically accurate region A in the Clarke error grid analyses.
Assuntos
Glicemia , Fotopletismografia , Pressão Sanguínea/fisiologia , Determinação da Pressão Arterial/métodos , Impedância Elétrica , Humanos , Redes Neurais de Computação , Fotopletismografia/métodosRESUMO
With the development of quantitative finance, machine learning methods used in the financial fields have been given significant attention among researchers, investors, and traders. However, in the field of stock index spot-futures arbitrage, relevant work is still rare. Furthermore, existing work is mostly retrospective, rather than anticipatory of arbitrage opportunities. To close the gap, this study uses machine learning approaches based on historical high-frequency data to forecast spot-futures arbitrage opportunities for the China Security Index (CSI) 300. Firstly, the possibility of spot-futures arbitrage opportunities is identified through econometric models. Then, Exchange-Traded-Fund (ETF)-based portfolios are built to fit the movements of CSI 300 with the least tracking errors. A strategy consisting of non-arbitrage intervals and unwinding timing indicators is derived and proven profitable in a back-test. In forecasting, four machine learning methods are adopted to predict the indicator we acquired, namely Least Absolute Shrinkage and Selection Operator (LASSO), Extreme Gradient Boosting (XGBoost), Back Propagation Neural Network (BPNN), and Long Short-Term Memory neural network (LSTM). The performance of each algorithm is compared from two perspectives. One is an error perspective based on the Root-Mean-Squared Error (RMSE), Mean Absolute Percentage Error (MAPE), and goodness of fit (R2). Another is a return perspective based on the trade yield and the number of arbitrage opportunities captured. Finally, a performance heterogeneity analysis is conducted based on the separation of bull and bear markets. The results show that LSTM outperforms all other algorithms over the entire time period, with an RMSE of 0.00813, MAPE of 0.70 percent, R2 of 92.09 percent, and an arbitrage return of 58.18 percent. Meanwhile, in certain market conditions, namely both the bull market and bear market separately with a shorter period, LASSO can outperform.
RESUMO
Rapid assessment of heavy metal (HM) pollution in mining areas is urgently required for further remediation. Here, hyperspectral technology was used to predict HM contents of multi-media environments (tailings, surrounding soils and agricultural soils) in a mining area. The correlation between hyperspectral data and HMs was explored, then the prediction models were established by partial least squares regression (PLSR) and back propagation neural networks (BPNN). The determination coefficients (R2), root mean squared error and ratios of performance to interquartile range (RPIQ) were used to evaluate the performance of the models. Results show that: (1) both PLSR and BPNN had good prediction ability, and (2) BPNN had better generalization ability (Cu (R2 = 0.89, RPIQ = 3.05), Sn (R2 = 0.86, RPIQ = 4.91), Zn (R2 = 0.74, RPIQ = 1.44) and Pb (R2 = 0.70, RPIQ = 2.10)). In summary, this study indicates that hyperspectral technology has potential application in HM estimation and soil pollution investigation in polymetallic mining areas.
Assuntos
Metais Pesados , Poluentes do Solo , China , Monitoramento Ambiental , Metais Pesados/análise , Solo , Poluentes do Solo/análise , EstanhoRESUMO
Oil-immersed transformers are expensive equipment in the electrical system, and their failure would lead to widespread blackouts and catastrophic economic losses. In this work, an elaborate diagnostic approach is proposed to evaluate twenty-six different transformers in-service to determine their operative status as per the IEC 60599:2022 standard and CIGRE brochure. The approach integrates dissolved gas analysis (DGA), transformer oil integrity analysis, visual inspections, and two Back Propagation Neural Network (BPNN) algorithms to predict the loss of life (LOL) of the transformers through condition monitoring of the cellulose paper. The first BPNN algorithm proposed is based on forecasting the degree of polymerization (DP) using 2-Furaldehyde (2FAL) concentration measured from oil samples using DGA, and the second BPNN algorithm proposed is based on forecasting transformer LOL using the 2FAL and DP data obtained from the first BPNN algorithm. The first algorithm produced a correlation coefficient of 0.970 when the DP was predicted using the 2FAL measured in oil and the second algorithm produced a correlation coefficient of 0.999 when the LOL was predicted using the 2FAL and DP output data obtained from the first algorithm. The results show that the BPNN can be utilized to forecast the DP and LOL of transformers in-service. Lastly, the results are used for hazard analysis and lifespan prediction based on the health index (HI) for each transformer to predict the expected years of service.
RESUMO
The use of the Storm Water Management Model (SWMM) to simulate flows in urban river watersheds necessitates the proper calibration of the various parameters involved in the process. Back Propagation Neural Network (BPNN) is often used to establish relationship between two sets of multivariate variables, such as parameters and simulation results of SWMM. The aim of this study is to establish an improved BPNN to calibrate SWMM. It was found that when using gauged flow data obtained from the urban river management system as calibration data, only using BPNN was not sufficient. An improved BPNN framework was proposed with integrating Principal Component Analysis (PCA) and Genetic Algorithm (GA) process, abbreviated as PCA-GA-BPNN. It was proved to be effective for calibration. The BPNN combined with GA process made 90 % of the predicted parameters within reasonable range, which was only 8 % using BPNN alone. The PCA process reduced the training time up to 64 %. Using a hydrograph of 196 h, compared with the nondominated sorting genetic algorithm (NSGA), PCA-GA-BPNN training time can be reduced from 18,142 s down to 4.5 s. Nash efficiency coefficients (NSE) of hydrographs fitting was 0.75. Including more rainfall events data in calibration achieved better fitting than including more gauging station data.
RESUMO
Atomic force microscopy (AFM) is a kind of high-precision instrument to measure the surface morphology of various conductive or nonconductive samples. However, obtaining a high-resolution image with standard AFM scanning requires more time. Using block compressive sensing (BCS) is an effective approach to achieve rapid AFM imaging. But, the routine BCS-AFM imaging is difficult to balance the image quality of each local area. It is easy to lead to excessive sampling in some flat areas, resulting in time-consuming. At the same time, there is a lack of sampling in some areas with significant details, resulting in poor imaging quality. Thus, an innovative adaptive BCS-AFM imaging method is proposed. The overlapped block is used to eliminate blocking artifacts. Characteristic parameters (GTV, Lu, and SD) are used to predict the local morphological characteristics of the samples. Back propagation neural network is employed to acquire the appropriate sampling rate of each sub-block. Sampling points are obtained by pre-scanning and adaptive supplementary scanning. Afterward, all sub-block images are reconstructed using the TVAL3 algorithm. Each sample is capable of achieving uniform, excellent image quality. Image visual effects and evaluation indicators (PSNR and SSIM) are employed for the purpose of evaluating and analyzing the imaging effects of samples. Compared with two nonadaptive and two other adaptive imaging schemes, our proposed scheme has the characteristics of a high degree of automation, uniformly high-quality imaging, and rapid imaging speed. HIGHLIGHTS: The proposed adaptive BCS method can address the issues of uneven image quality and slow imaging speed in AFM. The appropriate sampling rate of each sub-block of the sample can be obtained by BP neural network. The introduction of GTV, Lu, and SD can effectively reveal the morphological features of AFM images. Seven samples with different morphology are used to test the performance of the proposed adaptive algorithm. Practical experiments are carried out with two samples to verify the feasibility of the proposed adaptive algorithm.
RESUMO
The quantitative analysis of near-infrared spectroscopy in traditional Chinese medicine has still deficiencies in the selection of the measured indexes. Then Paeoniae Radix Alba is one of the famous "Eight Flavors of Zhejiang" herbs, however, it lacks the pharmacodynamic support, and cannot reflect the quality of Paeoniae Radix Alba accurately and reasonably. In this study, the spectrum-effect relationship of the anti-inflammatory activity of Paeoniae Radix Alba was established. Then based on the obtained bioactive component groups, the genetic algorithm, back propagation neural network, was combined with near-infrared spectroscopy to establish calibration models for the content of the bioactive components of Paeoniae Radix Alba. Finally, three bioactive components, paeoniflorin, 1,2,3,4,6-O-pentagalloylglucose, and benzoyl paeoniflorin, were successfully obtained. Their near-infrared spectroscopy content models were also established separately, and the validation sets results showed the coefficient of determination (R2 > 0.85), indicating that good calibration statistics were obtained for the prediction of key pharmacodynamic components. As a result, an integrated analytical method of spectrum-effect relationship combined with near-infrared spectroscopy and deep learning algorithm was first proposed to assess and control the quality of traditional Chinese medicine, which is the future development trend for the rapid inspection of traditional Chinese medicine.
Assuntos
Medicamentos de Ervas Chinesas , Espectroscopia de Luz Próxima ao Infravermelho , Medicamentos de Ervas Chinesas/farmacologia , Medicamentos de Ervas Chinesas/química , Controle de Qualidade , Redes Neurais de ComputaçãoRESUMO
Background: Research findings suggest that a significant proportion of individuals diagnosed with cancer, ranging from 25% to 60%, experience distress and require access to psycho-oncological services. Until now, only contemporary approaches, such as logistic regression, have been used to determine predictors of distress in oncological patients. To improve individual prediction accuracy, novel approaches are required. We aimed to establish a prediction model for distress in cancer patients based on a back propagation neural network (BPNN). Methods: Retrospective data was gathered from a cohort of 3063 oncological patients who received diagnoses and treatment spanning the years 2011-2019. The distress thermometer (DT) has been used as screening instrument. Potential predictors of distress were identified using logistic regression. Subsequently, a prediction model for distress was developed using BPNN. Results: Logistic regression identified 13 significant independent variables as predictors of distress, including emotional, physical and practical problems. Through repetitive data simulation processes, it was determined that a 3-layer BPNN with 8 neurons in the hidden layer demonstrates the highest level of accuracy as a prediction model. This model exhibits a sensitivity of 79.0%, specificity of 71.8%, positive predictive value of 78.9%, negative predictive value of 71.9%, and an overall coincidence rate of 75.9%. Conclusion: The final BPNN model serves as a compelling proof of concept for leveraging artificial intelligence in predicting distress and its associated risk factors in cancer patients. The final model exhibits a remarkable level of discrimination and feasibility, underscoring its potential for identifying patients vulnerable to distress.
RESUMO
BACKGROUND: Psycho-oncological support (PO) is an effective measure to reduce distress and improve the quality of life in patients with cancer. Currently, there are only a few studies investigating the (expressed) wish for PO. The aim of this study was to evaluate the number of patients who request PO and to identify predictors for the wish for PO. METHODS: Data from 3063 cancer patients who had been diagnosed and treated at a Comprehensive Cancer Center between 2011 and 2019 were analyzed retrospectively. Potential predictors for the wish for PO were identified using logistic regression. As a novelty, a Back Propagation Neural Network (BPNN) was applied to establish a prediction model for the wish for PO. RESULTS: In total, 1752 patients (57.19%) had a distress score above the cut-off and 14.59% expressed the wish for PO. Patients' requests for pastoral care (OR = 13.1) and social services support (OR = 5.4) were the strongest predictors of the wish for PO. Patients of the female sex or who had a current psychiatric diagnosis, opioid treatment and malignant neoplasms of the skin and the hematopoietic system also predicted the wish for PO, while malignant neoplasms of digestive organs and older age negatively predicted the wish for PO. These nine significant predictors were used as input variables for the BPNN model. BPNN computations indicated that a three-layer network with eight neurons in the hidden layer is the most precise prediction model. DISCUSSION: Our results suggest that the identification of predictors for the wish for PO might foster PO referrals and help cancer patients reduce barriers to expressing their wish for PO. Furthermore, the final BPNN prediction model demonstrates a high level of discrimination and might be easily implemented in the hospital information system.
RESUMO
AIM AND OBJECTIVE: Sulfonamides (sulfa drugs) are compounds with a wide range of biological activities and they are the basis of several groups of drugs. Quantitative Structure-Property Relationship (QSPR) models are derived to predict the logarithm of water/ 1-octanol partition coefficients (logP) of sulfa drugs. MATERIALS AND METHODS: A data set of 43 sulfa drugs was randomly divided into 3 groups: training, test and validation sets consisting of 70%, 15% and 15% of data point, respectively. A large number of molecular descriptors were calculated with Dragon software. The Genetic Algorithm - Multiple Linear Regressions (GA-MLR) and genetic algorithm -artificial neural network (GAANN) were employed to design the QSPR models. The possible molecular geometries of sulfa drugs were optimized at B3LYP/6-31G* level with Gaussian 98 software. The molecular descriptors derived from the Dragon software were used to build a predictive model for prediction logP of mentioned compounds. The Genetic Algorithm (GA) method was applied to select the most relevant molecular descriptors. RESULTS: The R2 and MSE values of the MLR model were calculated to be 0.312 and 5.074 respectively. R2 coefficients were 0.9869, 0.9944 and 0.9601for the training, test and validation sets of the ANN model, respectively. CONCLUSION: Comparison of the results revealed that the application the GA-ANN method gave better results than GA-MLR method.
Assuntos
Octanóis/química , Sulfonamidas/química , Água/química , Algoritmos , Difusão , Modelos Lineares , Modelos Químicos , Redes Neurais de Computação , Relação Quantitativa Estrutura-Atividade , Software , SolubilidadeRESUMO
The complex mixtures of local emission sources and regional transportations of air pollutants make accurate PM2.5 prediction a very challenging yet crucial task, especially under high pollution conditions. A symbolic representation of spatio-temporal PM2.5 features is the key to effective air pollution regulatory plans that notify the public to take necessary precautions against air pollution. The self-organizing map (SOM) can cluster high-dimensional datasets to form a meaningful topological map. This study implements the SOM to effectively extract and clearly distinguish the spatio-temporal features of long-term regional PM2.5 concentrations in a visible two-dimensional topological map. The spatial distribution of the configured topological map spans the long-term datasets of 25 monitoring stations in northern Taiwan using the Kriging method, and the temporal behavior of PM2.5 concentrations at various time scales (i.e., yearly, seasonal, and hourly) are explored in detail. Finally, we establish a machine learning model to predict PM2.5 concentrations for high pollution events. The analytical results indicate that: (1) high population density and heavy traffic load correspond to high PM2.5 concentrations; (2) the change of seasons brings obvious effects on PM2.5 concentration variation; and (3) the key input variables of the prediction model identified by the Gamma Test can improve model's reliability and accuracy for multi-step-ahead PM2.5 prediction. The results demonstrated that machine learning techniques can skillfully summarize and visibly present the clusted spatio-temporal PM2.5 features as well as improve air quality prediction accuracy.
RESUMO
Sports health literacy (SHL) is an important indicator of the all-round development of college students. However, the existing studies have not constructed a systematic and complete evaluation index system (EIS) or diversified the index weighting method. To solve the problem, this paper tries to evaluate and predict college students SHL based on artificial neural network (ANN). Firstly, an EIS was designed for college students SHL, including 4 goals, and 19 primary indices, and the structure of the evaluation and prediction system was presented for college students SHL. Next, college students SHL was comprehensively evaluated through analytic hierarchy process (AHP). Finally, a backpropagation neural network (BPNN) was established to predict college students SHL, and the initial weights were optimized by genetic algorithm (GA). The proposed EIS and prediction model were found scientific and effective through experiments. To sum up, the SHL of college students were evaluated and predicted in the following aspects: EIS construction, evaluation model establishment, and evaluation system design. The proposed model and system can comprehensively rate college students SHL. The statistical analysis of the ratings reflects the gaps between indices, and identifies those doing well and poorly on each index. Then, pertinent intervention can be implemented to satisfy the actual needs of improving college students SHL.(AU)
Assuntos
Humanos , Alfabetização , Estudantes , Redes Neurais de Computação , Esportes , Psicologia do EsporteRESUMO
This study introduced a patented novel methodological system for automatically analysis of Fourier Transform Infrared Spectrometer (FTIR) spectrum data located at 'fingerprint' region (wavenumber 670-800 cm(-1)), to simultaneously determinate multiple petroleum hydrocarbons (PHs) in real mixture samples. This system includes: an object oriented baseline correction; Band decomposition (curve fitting) method with mathematical optimization; and Artificial Neural Network (ANN) for determination, which is suitable for the characteristics of this IR regions, where the spectra are normally with low signal to noise ratio and high density of peaks. BTEX components are potentially lethal carcinogens and contained in many petroleum products. As a case study, six BTEX components were determinate automatically and simultaneously in mixture vapor samples. The robustness of the BTEX determination was validated using real petroleum samples, and the prediction results were compared with gas chromatography-mass spectrometry (GC-MS).
Assuntos
Derivados de Benzeno/análise , Espectroscopia de Infravermelho com Transformada de Fourier/métodos , Automação , Derivados de Benzeno/química , Filtração , Redes Neurais de Computação , Petróleo/análise , Fatores de Tempo , VolatilizaçãoRESUMO
Sodium potassium absorption ratio (SPAR) is an important measure of agricultural water quality, wherein four exchangeable cations (K(+), Na(+), Ca(2+) and Mg(2+)) should be simultaneously determined. An ISE-array is suitable for this application because its simplicity, rapid response characteristics and lower cost. However, cross-interferences caused by the poor selectivity of ISEs need to be overcome using multivariate chemometric methods. In this paper, a solid contact ISE array, based on a Prussian blue modified glassy carbon electrode (PB-GCE), was applied with a novel chemometric strategy. One of the most popular independent component analysis (ICA) methods, the fast fixed-point algorithm for ICA (fastICA), was implemented by the genetic algorithm (geneticICA) to avoid the local maxima problem commonly observed with fastICA. This geneticICA can be implemented as a data preprocessing method to improve the prediction accuracy of the Back-propagation neural network (BPNN). The ISE array system was validated using 20 real irrigation water samples from South Australia, and acceptable prediction accuracies were obtained.