Neural Network-based Regression Research Articles

Mapping the Continuous Cover of Invasive Noxious Weed Species Using Sentinel-2 Imagery and a Novel Convolutional Neural Regression Network

Mapping the Continuous Cover of Invasive Noxious Weed Species Using Sentinel-2 Imagery and a Novel Convolutional Neural Regression Network

An Artificial Neural Network-Based Fault Diagnostics Approach for Hydrogen-Fueled Micro Gas Turbines

The utilization of hydrogen fuel in gas turbines brings significant changes to the thermophysical properties of flue gas, including higher specific heat capacities and an enhanced steam content. Therefore, hydrogen-fueled gas turbines are susceptible to health degradation in the form of steam-induced corrosion and erosion in the hot gas path. In this context, the fault diagnosis of hydrogen-fueled gas turbines becomes indispensable. To the authors’ knowledge, there is a scarcity of fault diagnosis studies for retrofitted gas turbines considering hydrogen as a potential fuel. The present study, however, develops an artificial neural network (ANN)-based fault diagnosis model using the MATLAB environment. Prior to the fault detection, isolation, and identification modules, physics-based performance data of a 100 kW micro gas turbine (MGT) were synthesized using the GasTurb tool. An ANN-based classification algorithm showed a 96.2% classification accuracy for the fault detection and isolation. Moreover, the feedforward neural network-based regression algorithm showed quite good training, testing, and validation accuracies in terms of the root mean square error (RMSE). The study revealed that the presence of hydrogen-induced corrosion faults (both as a single corrosion fault or as simultaneous fouling and corrosion) led to false alarms, thereby prompting other incorrect faults during the fault detection and isolation modules. Additionally, the performance of the fault identification module for the hydrogen fuel scenario was found to be marginally lower than that of the natural gas case due to assumption of small magnitudes of faults arising from hydrogen-induced corrosion.

Deep Learning-based Lung dose Prediction Using Chest X-ray Images in Non-small Cell Lung Cancer Radiotherapy.

This study aimed to develop a deep learning model for the prediction of V20 (the volume of the lung parenchyma that received ≥20 Gy) during intensity-modulated radiation therapy using chest X-ray images. The study utilized 91 chest X-ray images of patients with lung cancer acquired routinely during the admission workup. The prescription dose for the planning target volume was 60 Gy in 30 fractions. A convolutional neural network-based regression model was developed to predict V20. To evaluate model performance, the coefficient of determination (R2), root mean square error (RMSE), and mean absolute error (MAE) were calculated with conducting a four-fold cross-validation method. The patient characteristics of the eligible data were treatment period (2018-2022) and V20 (19.3%; 4.9%-30.7%). The predictive results of the developed model for V20 were 0.16, 5.4%, and 4.5% for the R2, RMSE, and MAE, respectively. The median error was -1.8% (range, -13.0% to 9.2%). The Pearson correlation coefficient between the calculated and predicted V20 values was 0.40. As a binary classifier with V20 <20%, the model showed a sensitivity of 75.0%, specificity of 82.6%, diagnostic accuracy of 80.6%, and area under the receiver operator characteristic curve of 0.79. The proposed deep learning chest X-ray model can predict V20 and play an important role in the early determination of patient treatment strategies.

Attribute Subspace Partitioning with Neural Regression for Contextual Outlier Detection

Outlier detection is a critical machine learning task, especially when dealing with context-driven outliers. Such outliers usually remain undetected by conventional detection methods. Detecting these specific outliers requires deep domain knowledge. This paper introduces an innovative approach to contextual outlier detection, leveraging a hybrid method that combines a neural network-based regression model with traditional techniques. The proposed method enables precisely distinguishing contextual outliers by employing attribute subspace partitioning. The proposed algorithm effectively separates anomalous data points by integrating domain knowledge on attribute-specific contextual behavioral deviations with standard outlier detection methods. This experiment exhibits the effectiveness of the proposed method across five real-world datasets, yielding a notable enhancement in the range of 22% to 45% in the AUC score.

Neural network-based regression for heat transfer and fluid flow over in-line cylinder arrays with random pitch distances at low Reynolds number

Finding an arrangement, leading to a higher heat transfer and lower pressure drop, is crucial in the design of heat exchangers. Previous studies have primarily focused on regular arrangements with uniform pitch distances, which lack applicability to general configurations. In this study, we proposed a new procedure of a flow-learned building block (FLBB) to predict heat transfer in an in-line cylinder array with random pitch distances using a neural network-based regression analysis with a systematic data generation process. As a first step, we demonstrated the FLBB’s capability to predict the heat transfer and pressure drop in in-line cylinder arrays with random pitch distances at low Reynolds numbers from 1 to 100 for air ( ). Subsequently, a high-order FLBB approach was proposed to address the spatial interdependency between neighbouring cylinders, particularly in scenarios where vortex shedding occurs in the wake of cylinders at increased Reynolds numbers. The high-order FLBB approach was then shown to successfully describe various flow and temperature patterns using cylinder arrays with random pitch distances. The proposed procedure exhibited remarkable efficiency, requiring only about 1 s. Furthermore, the FLBB was successfully extended to various flow regimes, even encompassing unseen Reynolds numbers from 1 to 100.

Modeling of inversion layer capacitance of III-V double gate MOSFETs using a neural network-based regression technique

Modeling of inversion layer capacitance of III-V double gate MOSFETs using a neural network-based regression technique

On the impact of the galaxy window function on cosmological parameter estimation

ABSTRACT One important source of systematics in galaxy redshift surveys comes from the estimation of the galaxy window function. Up until now, the impact of the uncertainty in estimating the galaxy window function on parameter inference has not been properly studied. In this paper, we show that the uncertainty and the bias in estimating the galaxy window function will be salient for ongoing and next-generation galaxy surveys using a simulation-based approach. With a specific case study of cross-correlating emission-line galaxies from the DESI Legacy Imaging Surveys and the Planck cosmic microwave background lensing map, we show that neural network-based regression approaches to modelling the window function are superior in comparison to linear regression-based models. We additionally show that the definition of the galaxy overdensity estimator can impact the overall signal-to-noise of observed power spectra. Finally, we show that the additive biases coming from the window functions can significantly bias the modes of the inferred parameters and also degrade their precision. Thus, a careful understanding of the window functions will be essential to conduct cosmological experiments.

Deep learning regressor model based on nigrosome MRI in Parkinson syndrome effectively predicts striatal dopamine transporter-SPECT uptake

PurposeNigrosome imaging using susceptibility-weightedimaging (SWI) and dopamine transporter imaging using 123I-2β-carbomethoxy-3β-(4-iodophenyl)-N-(3-fluoropropyl)-nortropane (123I-FP-CIT) single-photon emission computerized tomography (SPECT) can evaluate Parkinsonism. Nigral hyperintensity from nigrosome-1 and striatal dopamine transporter uptake are reduced in Parkinsonism; however, quantification is only possible with SPECT. Here, we aimed to develop a deep-learning-based regressor model that can predict striatal 123I-FP-CIT uptake on nigrosome magnetic resonance imaging (MRI) as a biomarker for Parkinsonism.MethodsBetween February 2017 and December 2018, participants who underwent 3 T brain MRI including SWI and 123I-FP-CIT SPECT based on suspected Parkinsonism were included. Two neuroradiologists evaluated the nigral hyperintensity and annotated the centroids of nigrosome-1 structures. We used a convolutional neural network-based regression model to predict striatal specific binding ratios (SBRs) measured via SPECT using the cropped nigrosome images. The correlation between measured and predicted SBRs was evaluated.ResultsWe included 367 participants (203 women (55.3%); age, 69.0 ± 9.2 [range, 39–88] years). Random data from 293 participants (80%) were used for training. In the test set (74 participants [20%]), the measured and predicted 123I-FP-CIT SBRs were significantly lower with the loss of nigral hyperintensity (2.31 ± 0.85 vs. 2.44 ± 0.90) than with intact nigral hyperintensity (4.16 ± 1.24 vs. 4.21 ± 1.35, P < 0.01). The sorted measured 123I-FP-CIT SBRs and the corresponding predicted values were significantly and positively correlated (ρc = 0.7443; 95% confidence interval, 0.6216–0.8314; P < 0.01).ConclusionA deep learning-based regressor model effectively predicted striatal 123I-FP-CIT SBRs based on nigrosome MRI with high correlation using manually-measured values, enabling nigrosome MRI as a biomarker for nigrostriatal dopaminergic degeneration in Parkinsonism.

Continuous Non-Invasive Blood Pressure Measurement Using 60 GHz-Radar-A Feasibility Study.

Blood pressure monitoring is of paramount importance in the assessment of a human's cardiovascular health. The state-of-the-art method remains the usage of an upper-arm cuff sphygmomanometer. However, this device suffers from severe limitations-it only provides a static blood pressure value pair, is incapable of capturing blood pressure variations over time, is inaccurate, and causes discomfort upon use. This work presents a radar-based approach that utilizes the movement of the skin due to artery pulsation to extract pressure waves. From those waves, a set of 21 features was collected and used-together with the calibration parameters of age, gender, height, and weight-as input for a neural network-based regression model. After collecting data from 55 subjects from radar and a blood pressure reference device, we trained 126 networks to analyze the developed approach's predictive power. As a result, a very shallow network with just two hidden layers produced a systolic error of 9.2±8.3 mmHg (mean error ± standard deviation) and a diastolic error of 7.7±5.7 mmHg. While the trained model did not reach the requirements of the AAMI and BHS blood pressure measuring standards, optimizing network performance was not the goal of the proposed work. Still, the approach has displayed great potential in capturing blood pressure variation with the proposed features. The presented approach therefore shows great potential to be incorporated into wearable devices for continuous blood pressure monitoring for home use or screening applications, after improving this approach even further.

An ANN-assisted efficient enriched finite element method via the selective enrichment of moment fitting

Enrichment techniques that employ nonconforming mesh are effective in modeling structures with discontinuities because numerical issues regarding mesh quality are avoided. However, the accurate integration of the bilinear and linear forms on the discretized domain, which is required in the standard Galerkin-based finite element method, is computationally expensive due to the complexity of the enriched basis function. In this paper, we present a fast and accurate alternative method of numerical integration using nonlinear regression enabled by a multi-perceptron feedforward neural network. The relationship between an implicitly represented geometry and the quadrature rule derived from the moment fitting method is predicted by the neural network; the neural network-based regression model circumvents complex computation and significantly reduces the overall online time by avoiding expensive function evaluations. Through the selected numerical examples, we demonstrate the efficiency and accuracy of the current method, as well as the flexibility of the trained network to be used in different contexts.

A NN-based attention model for rope vibration prediction and channel sensitivity analysis

It is difficult to implement modeling for overall construction due to the complex structure and distinct environmental influences for the high-speed elevator so researchers rarely achieve end-to-end control for vibration suppression. Meanwhile, the dynamic parameters of the hoist component (the hoist rope) contain more environmental disturbance characteristics than the traction machine. In this paper, a sensitivity analysis of data channels of hoist rope that can help to clarify the effective control loop based on experimental results is implemented. The authors combined a neural network-based (NN-based) regression model and attention module for predicting hoisting rope vibration to achieve over 95% accuracy, and determine the channel data that contributes most to vibration prediction.

Neural network-based regression assisted PAPR reduction method for OFDM systems

Neural network-based regression assisted PAPR reduction method for OFDM systems

Neural Network-Augmented Locally Adaptive Linear Regression Model for Tabular Data

Creating an interpretable model with high predictive performance is crucial in eXplainable AI (XAI) field. We introduce an interpretable neural network-based regression model for tabular data in this study. Our proposed model uses ordinary least squares (OLS) regression as a base-learner, and we re-update the parameters of our base-learner by using neural networks, which is a meta-learner in our proposed model. The meta-learner updates the regression coefficients using the confidence interval formula. We extensively compared our proposed model to other benchmark approaches on public datasets for regression task. The results showed that our proposed neural network-based interpretable model showed outperformed results compared to the benchmark models. We also applied our proposed model to the synthetic data to measure model interpretability, and we showed that our proposed model can explain the correlation between input and output variables by approximating the local linear function for each point. In addition, we trained our model on the economic data to discover the correlation between the central bank policy rate and inflation over time. As a result, it is drawn that the effect of central bank policy rates on inflation tends to strengthen during a recession and weaken during an expansion. We also performed the analysis on CO2 emission data, and our model discovered some interesting explanations between input and target variables, such as a parabolic relationship between CO2 emissions and gross national product (GNP). Finally, these experiments showed that our proposed neural network-based interpretable model could be applicable for many real-world applications where data type is tabular and explainable models are required.

Development of Deep Learning Chest X-Ray Model for Cardiac Dose Prediction in Left-Sided Breast Cancer Radiotherapy

<h3>Purpose/Objective(s)</h3> Deep inspiration breath-hold (DIBH) is widely used to reduce the cardiac dose in left-sided breast cancer radiotherapy. This study aimed to develop a deep learning chest X-ray model for cardiac dose prediction to select patients with a potentially high risk of cardiac irradiation and need for DIBH radiotherapy. <h3>Materials/Methods</h3> We used 103 pairs of anteroposterior and lateral chest X-ray data of left-sided breast cancer patients (training cohort: n = 59, validation cohort: n=19, test cohort: n = 25). All patients underwent breast-conserving surgery followed by DIBH radiotherapy: the treatment plan consisted of two opposing tangential radiation fields. The prescription dose of the planning target volume was 42.56 Gy in 16 fractions. A convolutional neural network-based regression model was developed to predict the mean heart dose (∆MHD) reduction between free-breathing (MHD_FB) and DIBH. The model performance is evaluated as a binary classifier by setting the cutoff value of ∆MHD >1 Gy. <h3>Results</h3> The patient characteristics were as follows: the median (IQR) age was 52 (47–61) years, MHD_FB was 1.75 (1.14–2.47) Gy, and ∆MHD was 1.00 (0.52–1.64) Gy. The classification performance of the developed model showed a sensitivity of 85.7%, specificity of 90.9%, a positive predictive value of 92.3%, a negative predictive value of 83.3%, and a diagnostic accuracy of 88.0%. The AUC value of the ROC curve was 0.864. <h3>Conclusion</h3> The proposed model could predict ∆MHD in breast radiotherapy, suggesting the potential of a classifier in which patients are more desirable for DIBH.

Water Quality Retrieval from Landsat-9 (OLI-2) Imagery and Comparison to Sentinel-2

The Landsat series has marked the history of Earth observation by performing the longest continuous imaging program from space. The recent Landsat-9 carrying Operational Land Imager 2 (OLI-2) captures a higher dynamic range than sensors aboard Landsat-8 or Sentinel-2 (14-bit vs. 12-bit) that can potentially push forward the frontiers of aquatic remote sensing. This potential stems from the enhanced radiometric resolution of OLI-2, providing higher sensitivity over water bodies that are usually low-reflective. This study performs an initial assessment on retrieving water quality parameters from Landsat-9 imagery based on both physics-based and machine learning modeling. The concentration of chlorophyll-a (Chl-a) and total suspended matter (TSM) are retrieved based on physics-based inversion in four Italian lakes encompassing oligo to eutrophic conditions. A neural network-based regression model is also employed to derive Chl-a concentration in San Francisco Bay. We perform a consistency analysis between the constituents derived from Landsat-9 and near-simultaneous Sentinel-2 imagery. The Chl-a and TSM retrievals are validated using in situ matchups. The results indicate relatively high consistency among the water quality products derived from Landsat-9 and Sentinel-2. However, the Landsat-9 constituent maps show less grainy noise, and the matchup validation indicates relatively higher accuracies obtained from Landsat-9 (e.g., TSM R2 of 0.89) compared to Sentinel-2 (R2 = 0.71). The improved constituent retrieval from Landsat-9 can be attributed to the higher signal-to-noise (SNR) enabled by the wider dynamic range of OLI-2. We performed an image-based SNR estimation that confirms this assumption.

Development of deep learning chest X-ray model for cardiac dose prediction in left-sided breast cancer radiotherapy

Deep inspiration breath-hold (DIBH) is widely used to reduce the cardiac dose in left-sided breast cancer radiotherapy. This study aimed to develop a deep learning chest X-ray model for cardiac dose prediction to select patients with a potentially high risk of cardiac irradiation and need for DIBH radiotherapy. We used 103 pairs of anteroposterior and lateral chest X-ray data of left-sided breast cancer patients (training cohort: n = 59, validation cohort: n = 19, test cohort: n = 25). All patients underwent breast-conserving surgery followed by DIBH radiotherapy: the treatment plan consisted of three-dimensional, two opposing tangential radiation fields. The prescription dose of the planning target volume was 42.56 Gy in 16 fractions. A convolutional neural network-based regression model was developed to predict the mean heart dose (∆MHD) reduction between free-breathing (MHDFB) and DIBH. The model performance is evaluated as a binary classifier by setting the cutoff value of ∆MHD > 1 Gy. The patient characteristics were as follows: the median (IQR) age was 52 (47–61) years, MHDFB was 1.75 (1.14–2.47) Gy, and ∆MHD was 1.00 (0.52–1.64) Gy. The classification performance of the developed model showed a sensitivity of 85.7%, specificity of 90.9%, a positive predictive value of 92.3%, a negative predictive value of 83.3%, and a diagnostic accuracy of 88.0%. The AUC value of the ROC curve was 0.864. The proposed model could predict ∆MHD in breast radiotherapy, suggesting the potential of a classifier in which patients are more desirable for DIBH.

Design of a new optimized U-shaped lightweight liquid-cooled battery thermal management system for electric vehicles: A machine learning approach

Despite significant advances in battery thermal management system (BTMS), current electric vehicles (EVs) still lag behind user expectations in terms of battery performance and safety. A reliable, low-cost, and lightweight system is crucial for an effective BTMS that enables the battery's operation within a reasonable temperature range. In this study, a novel U-shaped lightweight liquid cooling method is introduced to optimize the thermal safety and weight of a prismatic battery cell. Numerical calculations are first carried out to investigate the effects of different fluid flow directions, flow rates, channel dimensions, fluid media, and discharge rates on the battery's temperature, and the results reveal that these factors simultaneously influence both the maximum temperature and temperature uniformity. To save computational costs and avoid the overwhelming complexity of the numerical measurements, a neural network-based regression model was constructed on the basis of 125 sets of simulation data, which comprised of four kernel functions with two outputs: maximum temperature and temperature difference. Each kernel function had a desirable testing accuracy (99.1%, 97.0%, 97.8%, 98.4%) after training, which indicates that the Gaussian Process Regression (GPR) with Matern 5/2 kernel function can better express complex non-linear relations for the prediction of maximum temperature and temperature difference with minimal errors between the simulated and predicted values compared with other kernel functions. Additionally, applying the GPR model to all cases studied, it is observed that Case 1 decreases the maximum temperature of the battery cell by 21% and lowers the weight of the cooling plate by 45% compared to the conventional liquid-cooled BTMS. Hence, it is chosen as the optimal design. The newly introduced BTMS can contribute to the development of lightweight and thermally safe battery cells.

Application of neural network based regression model to gas concentration analysis of TiO2 nanotube-type gas sensors

We performed a gas analysis of TiO2 nanotube (NT)-type integrated gas sensors using a machine learning (ML) algorithm and neural network-based regression. We fabricated a TiO2-NT integrated gas sensor with multiple sensing elements with different response characteristics, and we measured the output signals of each sensing element exposed to a gas mixture, where the main components were nitrogen and oxygen gas with a small amount of carbon monoxide. We analyzed the output signals of the sensor elements using the ML technique to predict the concentrations of CO and O2, to which the TiO2-NT gas sensors were sensitive. Sensor output data were collected for seven sets of mixed gas concentrations with different concentrations of each component gas. Four or five of the seven datasets were used as ML training data for the neural network method, and the concentrations of CO and O2 in the remaining three or two datasets were predicted. Consequently, we confirmed that increasing the number of sensor elements significantly improved the prediction accuracy of the gas concentration. When the output signals from 10 sensor elements were used, the gas concentration could be predicted with an accuracy of less than 0.001% for a carbon monoxide concentration of 0.02%. This accuracy was sufficient for practical application.

Vascular Aging Estimation Based on Artificial Neural Network Using Photoplethysmogram Waveform Decomposition: Retrospective Cohort Study.

BackgroundFor the noninvasive assessment of arterial stiffness, a well-known indicator of arterial aging, various features based on the photoplethysmogram and regression methods have been proposed. However, whether because of the existing characteristics not accurately reflecting the characteristics of the incident and reflected waveforms of the photoplethysmogram or because of the lack of expressive power of the regression model, a reliable arterial stiffness assessment technique based on a single photoplethysmogram has not yet been proposed.ObjectiveThe purpose of this study is to discover highly correlated features from the incident and reflected waves decomposed from a photoplethysmogram waveform and to develop an artificial neural network-based regression model for the assessment of vascular aging using newly derived features.MethodsWe obtained photoplethysmograms from 757 participants. All recorded photoplethysmograms were segmented for each beat, and each waveform was decomposed into incident and reflected waves by the Gaussian mixture model. The 26 basic features and 52 combined features were defined from the morphological characteristics of the incident and reflected waves. The regression model of the artificial neural network was developed using the defined features.ResultsIn correlation analysis, the features from the amplitude of the reflected wave and the skewness of the photoplethysmogram showed a relatively strong correlation with the participant’s real age. In the estimation of real age, the artificial neural network model showed 10.0 years of root mean square error. Its estimated age and real age had a strong correlation of 0.63 (P<.001).ConclusionsThis study proved that the features defined from the reflected wave and skewness of the photoplethysmogram are useful to assess vascular aging. Moreover, the regression model of artificial neural network using these features shows the feasibility for the estimation of vascular aging.

A generalized machine learning workflow to visualize mechanical discontinuity

Accurate detection and mapping of mechanical discontinuity in materials has widespread industrial and research applications. We developed a generalized machine-learning framework for visualizing single mechanical discontinuity embedded in material of any composition, velocity, density, porosity, and size with limited data. The proposed visualization of discontinuity requires accurate estimations of the length, location, and orientation of the embedded discontinuity by processing multipoint wave-transmission measurements. k-Wave simulator is used to create a large dataset of elastic waveforms recorded during multi-point wave-transmission measurements through materials containing single mechanical discontinuity. k-Wave simulator considers the wave attenuation, dispersion, and mode conversion in wave motion. Discrete wavelet transform (DWT) and statistical feature extraction are essential for data preprocessing prior to the data-driven model development. DWT also minimizes the effect of noise. Using hyper-parameter tuning and cross validation, gradient boosting regression can visualize the mechanical discontinuity with an accuracy of 0.85, in terms of coefficient of determination. A double-layered neural network-based regression has better performance with an accuracy of 0.95. Use of convolutional neural network converts the predictive task from a waveform processing to an image processing problem. Convolutional neural network achieved a generalization performance of 0.91. The proposed generalized workflow requires robust simulation of wave propagation, signal processing, feature engineering, and model evaluation. Sensors closest to the source and those located opposite the source are the most significant for the desired visualization. Notably, the sensors closest to the source capture the non-linear associations, whereas the sensor on the border opposite to the source capture the linear associations between the measured waveforms and the properties of the mechanical discontinuity.