Normalized Root Mean Square Error Research Articles

Quantitative Relationship of Plant Height and Leaf Area Index of Spring Maize under Different Water and Nitrogen Treatments Based on Effective Accumulated Temperature

To optimize the growth management of spring maize, it is essential to understand the dynamics of plant height and leaf area index (LAI) under controlled water and nitrogen supply. This study conducted two-year field experiments (2022–2023) in Karamay, Xinjiang. Three irrigation levels (75%, 100%, and 125% of Crop Evapotranspiration (ETc)) and four nitrogen application rates (0, 93, 186, and 279 kg N/ha) were set. A logistic growth model was fitted using accumulated effective temperature as the independent variable to analyze the growth and development characteristics of spring maize under various water and nitrogen conditions. The results demonstrated that the logistic models, based on relative effective accumulated temperature, had a determination coefficient (R2) of over 0.99 and a Normalized Root Mean Square Error (NRMSE) of less than 10%. Irrigation extended the rapid growth phase of plant height, whereas nitrogen application shortened the time to enter this rapid growth phase and prolonged its duration. Irrigation increased the maximum LAI growth rate and shortened and prolonged the rapid growth phase, while nitrogen extended the duration of the rapid growth phase for LAI. The W2N2 treatment, consisting of 100% ETc irrigation and 186 kg N/ha, was identified as the optimal drip irrigation water–nitrogen combination for spring maize in the study area. Under optimal water and nitrogen supply, both the maximum growth rate and the average growth rate during the rapid growth phase were higher, requiring accumulated effective temperatures of 825.16–845.74 °C·d and 856.68–890.00 °C·d, respectively, to reach these rates. The appropriate water and nitrogen supply significantly enhanced the synergistic promotion of growth and development in spring maize. This study provides a theoretical basis for the quantitative analysis of growth dynamics in summer maize using effective accumulated temperature.

Using Transfer Learning and XGBoost for Early Detection of Fires in Offshore Wind Turbine Units

To improve the power generation efficiency of offshore wind turbines and address the problem of high fire monitoring and warning costs, we propose a data-driven fire warning method based on transfer learning for wind turbines in this paper. This paper processes wind turbine operation data in a SCADA system. It uses an extreme gradient-boosting tree (XGBoost) algorithm to build an offshore wind turbine unit fire warning model with a multiparameter prediction function. This paper selects some parameters from the dataset as input variables for the model, with average cabin temperature, average outdoor temperature, average cabin humidity, and average atmospheric humidity as output variables. This paper analyzes the distribution information of input and output variables and their correlation, analyzes the predicted difference, and then provides an early warning for wind turbine fires. This paper uses this fire warning model to transfer learning to different models of offshore wind turbines in the same wind farm to achieve fire warning. The experimental results show that the prediction performance of the multiparameter is accurate, with an average MAPE of 0.016 and an average RMSE of 0.795. It is better than the average MAPE (0.051) and the average RMSE (2.020) of the prediction performance of a backpropagation (BP) neural network, as well as the average MAPE (0.030) and the average RMSE (1.301) of the prediction performance of random forest. The transfer learning model has good prediction performance, with an average MAPE of 0.022 and an average RMSE of 1.469.

Development and validation of a sunflower crop growth module for the Daisy model

The objective of this study was to develop a crop model for sunflower in Daisy. For this intent, the field experiment data of two full irrigation treatments for 2 years, 2012 and 2013, were used to develop the crop model. The average RMSE of the dry matter (DM), leaf area index (LAI), height of plant (H), and water content parameters in the treatments utilized for calibration was calculated 0.093 Ton/ha, 0.198 cm2/cm2, 2.25 cm, and 1.035%, respectively. The average RMSE of the parameters of DM, LAI, H, and water content in the treatments utilized for validation were estimated 0.207 Ton/ha, 0.272 cm2/cm2, 4.22 cm, and 1.123%, respectively. The results of the validation showed that the developed crop model had good accuracy making it useful for simulation of sunflower agroecology. It is suggested to use the developed crop model in the information library of Daisy and survey and calibrate it for other areas with different climates and sunflower varieties.

Sound pressure level determination around the transformer using coherent and incoherent analytical summation methods

In this paper, coherent and incoherent analytical summation methods from elementary sound sources on the transformer’s tank are presented. They aim to determine a one-third-octave spectrum of sound pressure levels (SPL) in points around the transformer. Each part of the transformer’s surface is treated as a separate plane source so the analytical calculation can be applied to find SPL in the surroundings. Reflections from the ground are also considered by putting an image of the plane sources on the other side of the reflecting plane and adding its contributions to the total pressure in the desired points. Grid-like vibration measurements of vibration velocity on the tank surfaces are used as input parameters. Vibrations and sound pressure levels are measured to validate the method on the 5 MVA transformer experimental object. The SPL around the transformer in short-circuit (SC) and open-circuit (OC) tests is measured in the semi-anechoic chamber to compare it with theoretical results. By analyzing the results, the coherent calculation with reflection provided the most accurate results. In the SC operating condition, the normalized root mean square error (NRMSE) is 17.2%, and in the OC operating condition, it is 10.9%. The novelty of the presented method is that it considers complicated transformer geometry where each surface is calculated as a separate noise source. It calculates noise at a distance from the elementary vibration sources, considering phase and reflection from the hard ground, and provides detailed noise maps that can be used for noise modeling around the substations.

Models to predict nitrogen excretion from beef cattle fed a wide range of diets compiled from South America.

The objective of this meta-analysis was to develop and evaluate models for predicting nitrogen (N) excretion in feces, urine, and manure in beef cattle in South America. The study incorporated a total of 1,116 individual observations of N excretion in feces and 939 individual observations of N excretion in feces and in urine (g/d), representing a diverse range of diets, animal genotypes, and management conditions in South America. The datasetalso included data on dry matter intake (DMI; kg/d) and nitrogen intake (NI; g/d), concentrations of dietary components, as well as average daily gain (ADG; g/d) and average body weight (BW; kg). Models were derived using linear mixed-effects regression with a random intercept for the study. Fecal N excretion was positively associated with DMI, NI, nonfibrous carbohydrates, average BW, and ADG and negatively associated with EE and CP concentration in the diet. The univariate model predicting fecal N excretion based on DMI (model 1) performed slightly better than the univariate model, which used NI as a predictor variable (model 2) with a root mean square error (RMSE) of 38.0 vs. 39.2%, the RMSE-observations SD ratio (RSR) of 0.81 vs. 0.84, and concordance correlation coefficient (CCC) of 0.53 vs. 0.50, respectively. Models predicting urinary N excretion were less accurate than those derived to predict fecal N excretion, with an average RMSE of 43.7% vs. 37.0%, respectively. Urinary and manure N excretion were positively associated with DMI, NI, CP, average BW, and ADG and negatively associated with neutral detergent fiber concentration in the diet. As opposed to fecal N excretion, the univariate model predicting urinary N excretion using NI (model 10) performed slightly better than the univariate model using DMI (model 9) as predictor variable with an RMSE of 36.0% vs. 39.7%, RSR 0.85 vs. 0.93, and CCC of 0.43 vs. 0.29, respectively. The models developed in this study are applicable for predicting N excretion in beef cattle across a broad spectrum of dietary compositions and animal genotypes in South America. The univariate model using DMI as a predictor is recommended for fecal N prediction, while the univariate model using NI is recommended for predicting urinary and manure N excretion because the use of more complex models resulted in little to no benefits. However, it may be more useful to consider more complex models that incorporate nutrient intakes and diet composition for decision-making when N excretion is a factor to be considered. Three extant equations evaluated in this study have the potential to be used in tropical conditions typical of South America to predict fecal N excretion with good precision and accuracy. However, none of the extant equations are recommended for predicting urine or manure N excretion because of their high RMSE, and low precision and accuracy.

Wind Shear Model Considering Atmospheric Stability to Improve Accuracy of Wind Resource Assessment

An accurate wind shear model is an important prerequisite in extrapolating the wind resource from lower heights to the increasing hub height of wind turbines. Based on the 1-year dataset (collected in 2014) consisting of 15-minute intervals collected at heights of 2, 10, 50, 100, and 150 m on an anemometer tower in northern China, the present study focuses on the time-varying relationship between the wind shear coefficient (WSC) and atmospheric stability and proposes a wind shear model considering atmospheric stability. Through the relationship between Monin–Obukhov (M-O) length and gradient Richardson number, the M-O length is directly calculated by wind data, and the WSC is calculated by combining the Panofsky and Dutton (PD) models, which enhances the engineering practicability of the model. Then, the performance of the model is quantified and compared with two alternative methods: the use of annual average WSC and the use of stability change WSC extrapolation. The analysis demonstrates that the proposed model outperforms the other approaches in terms of normal root mean square error (NRMSE) and normal bias (NB). More specifically, this method reduces the NRMSE and NB by 24–29% and 76–95%, respectively. Meanwhile, it reaches the highest extrapolation accuracy under unstable and stable atmospheric conditions. The results are verified using the Weibull distribution.

Assessing the long‐term effects of conservation agriculture on cotton production in Northeast Louisiana using the denitrification–decomposition model

AbstractConservation agriculture (CA) aims to sustain agricultural production, soil, and environmental health in agroecosystems and has been promoted throughout the United States. The adoption of CA in cotton (Gossypium hirsutum) systems provides both agronomic and environmental benefits. Yet, there is limited information on the long‐term effects of CA practices on crop yield and adaptation strategies. An integrated CA system, that is, cover crops with no‐tillage (NT) instead of conventional agriculture, was implemented in the long‐term field experiments and assessed with an integrated biogeochemical model. Using the denitrification–decomposition model, this study estimated the effects of four different cover crops, for example, native grass (NG), hairy vetch (Vicia villosa), winter wheat (Triticum aestivum L.), and crimson clover (Trifolium incarnatum), on cotton yield under four different nitrogen (N) levels (e.g., 0, 50, 100, and 150 kg N/ha) and estimated responses on carbon (C) sequestration, and ecosystem functionality over a 10‐year study. The NT‐NG 50 N was used as a calibration dataset to accurately estimate the cotton lint yield with a normalized root mean square error (NRMSE) of 21% and model efficiency of 0.3. The calibration data validated the effects of hairy vetch, winter wheat, and crimson clover under the NT‐50 N with NRMSE of 24%, 21%, and 25%, respectively. According to the scenario analysis, the 50 kg N/ha application with a single‐irrigation event (10‐cm depth) was most beneficial for maximizing the cotton yield with cover crop incorporation at the NT system over the long term. The effects of increasing cover crop biomass (i.e., double seed rate) on C content, regardless of N application rates, varied based on the relationship between the main and cover crop species. Besides, the furrow plow tillage system provided efficient C sequestration. The proposed approach stands to provide agricultural and environmental sustainability with the implementation of cover crop or crop residue incorporation instead of increased N application, seed rates, and irrigation events under NT practices.

Deep learning generation of preclinical positron emission tomography (PET) images from low-count PET with task-based performance assessment.

Preclinical low-count positron emission tomography (LC-PET) imaging offers numerous advantages such as facilitating imaging logistics, enabling longitudinal studies of long- and short-lived isotopes as well as increasing scanner throughput. However, LC-PET is characterized by reduced photon-count levels resulting in low signal-to-noise ratio (SNR), segmentation difficulties, and quantification uncertainties. We developed and evaluated a novel deep-learning (DL) architecture-Attention based Residual-Dilated Net (ARD-Net)-to generate standard-count PET (SC-PET) images from LC-PET images. The performance of the ARD-Net framework was evaluated for numerous low count realizations using fidelity-based qualitative metrics, task-based segmentation, and quantitative metrics. Patient Derived tumor Xenograft (PDX) with tumors implanted in the mammary fat-pad were subjected to preclinical [18F]-Fluorodeoxyglucose (FDG)-PET/CT imaging. SC-PET images were derived from a 10min static FDG-PET acquisition, 50min post administration of FDG, and were resampled to generate four distinct LC-PET realizations corresponding to 10%, 5%, 1.6%, and 0.8% of SC-PET count-level. ARD-Net was trained and optimized using 48 preclinical FDG-PET datasets, while 16 datasets were utilized to assess performance. Further, the performance of ARD-Net was benchmarked against two leading DL-based methods (Residual UNet, RU-Net; and Dilated Network, D-Net) and non-DL methods (Non-Local Means, NLM; and Block Matching 3D Filtering, BM3D). The performance of the framework was evaluated using traditional fidelity-based image quality metrics such as Structural Similarity Index Metric (SSIM) and Normalized Root Mean Square Error (NRMSE), as well as human observer-based tumor segmentation performance (Dice Score and volume bias) and quantitative analysis of Standardized Uptake Value (SUV) measurements. Additionally, radiomics-derived features were utilized as a measure of quality assurance (QA) in comparison to true SC-PET. Finally, a performance ensemble score (EPS) was developed by integrating fidelity-based and task-based metrics. Concordance Correlation Coefficient (CCC) was utilized to determine concordance between measures. The non-parametric Friedman Test with Bonferroni correction was used to compare the performance of ARD-Net against benchmarked methods with significance at adjusted p-value ≤0.01. ARD-Net-generated SC-PET images exhibited significantly better (p ≤ 0.01 post Bonferroni correction) overall image fidelity scores in terms of SSIM and NRMSE at majority of photon-count levels compared to benchmarked DL and non-DL methods. In terms of task-based quantitative accuracy evaluated by SUVMeanand SUVPeak, ARD-Net exhibited less than 5% median absolute bias for SUVMean compared to true SC-PET and lower degree of variability compared to benchmarked DL and non-DL based methods in generating SC-PET. Additionally, ARD-Net-generated SC-PET images displayed higher degree of concordance to SC-PET images in terms of radiomics features compared to non-DL and other DL approaches. Finally, the ensemble score suggested that ARD-Net exhibited significantly superior performance compared to benchmarked algorithms (p ≤ 0.01 post Bonferroni correction). ARD-Net provides a robust framework to generate SC-PET from LC-PET images. ARD-Net generated SC-PET images exhibited superior performance compared other DL and non-DL approaches in terms of image-fidelity based metrics, task-based segmentation metrics, and minimal bias in terms of task-based quantification performance for preclinical PET imaging.

Dingo optimization algorithm-based random forests model to evaluate the compressive strength of the concrete at elevated temperatures

The qualities of the materials employed to manufacture concrete are significantly impacted by high temperatures, which results in a noticeable decrease in the material’s strength characteristics. Concrete must be worked very hard and allowed to reach the required compressive strength (fc). Nevertheless, a preliminary estimation of the desired outcome may be made with an outstanding degree of reliability by using supervised machine learning algorithms. The study combined the Dingo optimization algorithm (DOA), Coot bird optimization (COA), and Artificial rabbit optimization (ARO) with Random Forests (RF) evaluation to determine the fc of concrete at high temperatures. The abbreviations used for the combined methods are RFD, RFC, and RFA, respectively. Remarkably, removing the temperature (T) parameter from the input set leads to a remarkable 1100% improvement in the effectiveness index (PI) and normalized root mean squared error (NRMSE), while causing a significant fall in the coefficient of determination (R2). The findings suggest that all RFD, RFC, and RFA have substantial promise in properly forecasting the fc of concrete at high temperatures. More precisely, the RFD algorithm demonstrated exceptional precision with R2 values of 0.9885 and 0.9873 throughout the training and testing stages, respectively. Through a comparison of the error percentages for RFD, RFC, and RFA in error-based measurements, it becomes evident that RFD exhibits an error rate that is about 50% smaller compared to that of RFC and RFA. This prediction is crucial for various industries and applications where concrete structures are subjected to elevated temperatures, such as in fire resistance assessments for buildings, tunnels, bridges, and other infrastructure. By accurately forecasting the compressive strength of concrete under these conditions, engineers and designers can make informed decisions regarding the material’s suitability and performance in high-temperature environments, leading to enhanced safety, durability, and cost-effectiveness of structures.

Satellite-based soybean yield prediction in Argentina: A comparison between panel regression and deep learning methods

The accurate prediction of soybean yield is vital for global food market stabilization and food security. Recent advancements in remote sensing technology have significantly amplified interest in leveraging satellite-based methods for predicting crop yield. These methods offer in-season yield estimates. By utilizing this timely information, decision-makers can formulate strategic, well-informed choices that preemptively mitigate potential food price hikes, ultimately bolstering food security. While simple regression models have been widely utilized for satellite-based yield prediction, researchers have recently begun to explore the use of deep learning algorithms. This study compares the performance of panel regression and deep learning models for in-season soybean yield prediction at the Department (county-equivalent) level in Argentina. Data sources include the latest soybean land use products and MODIS bi-weekly vegetation index products. Results indicate that deep learning models significantly outperform panel regression. Deep learning Long Short-Term Memory (LSTM) models, which incorporate attention mechanism and a series of peak NDVI images, generate more accurate and time-sensitive predictions. Among competing LSTM models, the one with attention mechanism applied to the entire growing season's NDVI data yields the highest prediction accuracy, with a Root Mean Square Error (RMSE) of 505.78 kg/ha and Normalized Root Mean Square Error (NRMSE) of 0.0726. The LSTM model with attention on the three highest NDVI images attains a satisfactory prediction accuracy (RMSE = 627.28 kg/ha, NRMSE = 0.089) six weeks prior to harvest. This study presents a robust model for predicting crop yields, promoting sustainable production of soybeans and facilitating knowledgeable choices among farmers and policymakers.

Abstract PO5-07-03: Integrating mechanism-based and data-driven modeling to predict the response of triple negative breast cancer to therapy

Abstract Introduction Neoadjuvant chemotherapy (NAC) is the standard-of-care (SOC) for patients with locally advanced triple-negative breast cancer (TNBC) [1]. Unfortunately, only about half of TNBC patients achieve a pathological complete response (pCR) at the completion of NAC [2]. With the recent FDA approval of immunotherapy, the response rate improved by about 8% [3]. To try to further improve NAC response rates, we build upon methods that predict treatment response to optimize interventions and outcomes. Our previous studies have demonstrated that excellent predictive accuracy of TNBC response to NAC was achieved through calibrating mechanism-based models to both pre- and on- treatment imaging data [4]. However, by requiring on-treatment imaging data, therapy optimization can only occur after a patient begins NAC. Here, our goal is to relax the requirement of on-treatment imaging data such that we can predict, and potentially optimize, a patient’s response to NAC before initiating it. We will accomplish this through combining data-driven approaches with mechanism-based modeling to obtain spatiotemporal predictions of tumor response prior to NAC. More specifically, we integrate a reaction-diffusion equation based mathematical model with a U-Net based convolutional neural network (CNN). Methods The reaction-diffusion equation models the development of tumor cellularity over time as the sum of tumor cell diffusion, controlled by a fixed diffusivity, and tumor cell proliferation, controlled by a calibrated net proliferation rate. The net proliferation rate implicitly accounts for both proliferative and drug induced death effects. We calibrated the net proliferation rates in this model on a patient-specific basis for 128 patients from the ARTEMIS trial (NCT02276443) using three imaging timepoints – pre- (V1), post two cycles (V2), and post four cycles (V3) of A/C. We then employed a CNN to characterize the relationship between the calibrated net proliferation rates and the pre-treatment imaging data. The CNN takes pretreatment imaging inputs of the apparent diffusion coefficient map, percent enhancement maps from the dynamic contrast enhanced MRI timecourse, and precontrast T1 map. The CNN outputs an estimate of the calibrated net proliferation rate which can be used to run the mathematical model forward in time to predict cellularity maps at V2/3. In training the CNN, our loss considers both the concordance correlation coefficient (CCC) and the normalized root mean square error (NRMSE). We calculated loss on the net proliferation rate predictions as NRMSE with a CCC penalty to aid in characterizing the non-Gaussian distribution of the net proliferation rate. We calculated loss on the cellularity predictions as the CCC between predicted and measured voxel-wise cellularity. The CNN was trained using the Adam optimizer and a fivefold cross validation with a withheld test set of 26 patients. Results Using the CNN predictions to forward run the mechanism-based model in time to V3 we calculated change in total tumor cellularity and volume from V1 to V3. In the test cohort, we obtained CCC values between the predicted and measured changes of 0.95 and 0.91 for cellularity and volume, respectively. Discussion and summary Through integrating mechanistic modeling and deep learning, we can accurately predict the spatiotemporal development of TNBC response to NAC using only pretreatment imaging data. We will build upon these results by expanding our model to include separate proliferation and drug induced death terms to allow for therapy intervention and optimization. Additionally, we will incorporate other pretreatment data types including genetic data. [1] Liedtke et al., J of Clin Oncol, 2008. [2] Spring et al., Clin Cancer Res, 2020. [3] Schmid et al., NEJM, 2020. [4] Wu et al., Cancer Res, 2022. Citation Format: Casey Stowers, Chengyue Wu, Sidharth Kumar, Zhan Xu, Clinton Yam, Jong Bum Son, Jingfei Ma, Jonathan Tamir, Gaiane Rauch, Thomas Yankeelov. Integrating mechanism-based and data-driven modeling to predict the response of triple negative breast cancer to therapy [abstract]. In: Proceedings of the 2023 San Antonio Breast Cancer Symposium; 2023 Dec 5-9; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2024;84(9 Suppl):Abstract nr PO5-07-03.

Swin Transformer and the Unet Architecture to Correct Motion Artifacts in Magnetic Resonance Image Reconstruction.

We present a deep learning-based method that corrects motion artifacts and thus accelerates data acquisition and reconstruction of magnetic resonance images. The novel model, the Motion Artifact Correction by Swin Network (MACS-Net), uses a Swin transformer layer as the fundamental block and the Unet architecture as the neural network backbone. We employ a hierarchical transformer with shifted windows to extract multiscale contextual features during encoding. A new dual upsampling technique is employed to enhance the spatial resolutions of feature maps in the Swin transformer-based decoder layer. A raw magnetic resonance imaging dataset is used for network training and testing; the data contain various motion artifacts with ground truth images of the same subjects. The results were compared to six state-of-the-art MRI image motion correction methods using two types of motions. When motions were brief (within 5 s), the method reduced the average normalized root mean square error (NRMSE) from 45.25% to 17.51%, increased the mean structural similarity index measure (SSIM) from 79.43% to 91.72%, and increased the peak signal-to-noise ratio (PSNR) from 18.24 to 26.57 dB. Similarly, when motions were extended from 5 to 10 s, our approach decreased the average NRMSE from 60.30% to 21.04%, improved the mean SSIM from 33.86% to 90.33%, and increased the PSNR from 15.64 to 24.99 dB. The anatomical structures of the corrected images and the motion-free brain data were similar.

Kalman filter based on a fractional discrete-time stochastic augmented CoVid-19 model

In this paper, we study the dynamics of the CoVid-19 outbreak in Semarang, Indonesia, using a fractional CoVid-19 model. We first determine the effects of the isolation rate ∊ and infection rate β on the reproduction number R0 and infected number V. We find that R0 is directly proportional to β and inversely proportional to ∊. For V, the effect of physical distancing is not as significant as changing ∊. As ∊ increases, V decreases, the number of susceptible individuals increases, the number of quarantined individuals decreases sharply, and the number of recovered individuals decreases. Moreover, the effect of vaccination is also considered. The combination of physical distancing, isolation, and vaccination has a significant impact on reducing the number of infected individuals. Analysis of dynamical systems allows us to understand the characteristics of our model, such as its boundedness and non-negativity, the existence of equilibrium points, the existence and uniqueness of solutions, and the local and global stability. To validate our fractional CoVid-19 model, we introduce the fractional extended Kalman filter (FracEKF) as a prediction method and compare the results against reported CoVid-19 data. FracEKF is a modified version of the basic extended Kalman filter with a time-fractional memory effect. The prediction results illustrate the accuracy of this model in terms of the root mean square error (RMSE), normalized root mean square error (NRMSE), and mean absolute percentage error (MAPE) for each fractional-order. Varying ∊ reproduces the trends observed in the reported data for the number of infected individuals, i.e., when ∊ increases, the infected number decreases. Moreover, a higher fractional-order results in higher model accuracy. Furthermore, higher values of the process noise Qf give smaller errors, whereas higher values of the observation noise Rf produce higher errors. Qf and the fractional-order α are inversely proportional to RMSE,NRMSE, and MAPE, whereas Rf is directly proportional to RMSE,NRMSE, and MAPE.

Estimating soil moisture content in citrus orchards using multi-temporal sentinel-1A data-based LSTM and PSO-LSTM models

Soil moisture content is a vital variable in agricultural, hydrological, ecological and climatological processes. However, susceptible to soil type, soil structure, topography, vegetation and human activities, soil moisture content exhibits strong spatial heterogeneity in spatial distribution, which makes it difficult to accurately estimate the soil moisture content distribution information at a large scale using conventional methods. To solve the problem, this study proposed a novel hybrid model (PSO-LSTM) based on the particle swarm optimization (PSO) and long short-term memory (LSTM) network model to accurately predict soil moisture content at a large scale. Five different input combinations were constructed based on the vertical polarization (VV) and cross-polarization (VH) of multi-phase Sentinel-1A data, and the soil moisture content at depths of 5 cm, 10 cm, 20 cm and 40 cm in citrus orchards were estimated using the standalone LSTM and hybrid PSO-LSTM models. The results showed that the estimation accuracy of the hybrid PSO-LSTM model was greater than that of the standalone LSTM model at different depths, with the normalized root mean square error (NRMSE) of 4.568–11.023 % and 18.056–30.156 %, respectively. With the VV polarization as the only inputs, the PSO-LSTM model obtained high prediction accuracy, with the normalized root mean square error (NRMSE) of 5.458–10.125 %, respectively. Therefore, the PSO-LSTM model with VV polarization input was recommended to estimate the soil moisture content at different depths in citrus orchards, which provides important data for decision-making on distributed precision irrigation at a large scale.

Predicting the performance of a photovoltaic unit via machine learning methods in the existence of finned thermal storage unit

The process of cooling the photovoltaic (PV) cells is essential since increasing solar cell temperatures reduces electrical effectiveness. One suggested method to reduce cell temperature is employing PCM (phase change material) mixed with nanomaterial. Thermoelectric modules between the layers are also recommended to increase electrical efficiency. By taking advantage of machine learning predictive algorithms, the objective is to forecast the effectiveness of a PV unit equipped with finned thermal storage unit in existence of nanomaterials, while maintaining an appropriate accuracy based on two evaluation metrics throughout prediction. According to the data analytics of the simulation, the selected models used to achieve this goal are linear regression, polynomial regression, lasso regression, and an Auto-Regressive Integrated Moving Average (ARIMA) model. These models were then examined employing the root mean squared error (RMSE) and mean absolute percentage error (MAPE) across three cases, namely plain mode, mode with fins, and plain mode with the effect of gravity. The test RMSE and MAPE of these models across the above-mentioned cases were then calculated for comparison purposes, where the best-performing model for predicting both PCM temperature and melting fraction (MF) was ARIMA, with an average RMSE of 0.11, and an average MAPE of 0.0003 for predicting PCM temperature and an average RMSE of 0.003, and an average MAPE of 0.0034 for predicting MF across all cases. The ARIMA model was then used to predict the MF, PCM temperature, and PV temperature of the last case, namely mode with fins and the effect of gravity. Finally, this model was also used to predict the time at which the PCMs are completely melted in all modes, as well as to predict PV temperature. The total electrical efficiency of the mode with fins and the effect of gravity were calculated using the predicted PV temperature, yielding a result of 13.993 %.

FlexDTI: flexible diffusion gradient encoding scheme-based highly efficient diffusion tensor imaging using deep learning.


Most deep neural network-based diffusion tensor imaging methods require the diffusion gradients' number and directions in the data to be reconstructed to match those in the training data. This work aims to develop and evaluate a novel dynamic-convolution-based method called FlexDTI for highly efficient diffusion tensor reconstruction with flexible diffusion encoding gradient scheme.
Approach:
FlexDTI was developed to achieve high-quality DTI parametric mapping with flexible number and directions of diffusion encoding gradients. The method used dynamic convolution kernels to embed diffusion gradient direction information into feature maps of the corresponding diffusion signal. Furthermore, it realized the generalization of a flexible number of diffusion gradient directions by setting the maximum number of input channels of the network. The network was trained and tested using datasets from the Human Connectome Project and local hospitals. Results from FlexDTI and other advanced tensor parameter estimation methods were compared.
Main results:
Compared to other methods, FlexDTI successfully achieves high-quality diffusion tensor-derived parameters even if the number and directions of diffusion encoding gradients change. It reduces normalized root mean squared error (NRMSE) by about 50% on fractional anisotropy (FA) and 15% on mean diffusivity (MD), compared with the state-of-the-art deep learning method with flexible diffusion encoding gradient scheme.
Significance:
FlexDTI can well learn diffusion gradient direction information to achieve generalized DTI reconstruction with flexible diffusion gradient scheme. Both flexibility and reconstruction quality can be taken into account in this network.&#xD.

A Building Heat Load Prediction Method Driven by a Multi-Component Fusion LSTM Ridge Regression Ensemble Model

Under the background of “double carbon”, building carbon emission reduction is urgent, and improving energy efficiency through short-term building heat load forecasting is an efficient means of building carbon emission reduction. Aiming at the characteristics of the decomposed short-term building heat load data, such as complex trend changes, significant seasonal changes, and randomness, a single-step short-term building heat load prediction method driven by the multi-component fusion LSTM Ridge Regression Ensemble Model (ST-LSTM-RR) is designed and implemented. First, the trend and seasonal components of the heat load are decomposed by the STL seasonal decomposition algorithm, which are fused into the original data to construct three diversified datasets; second, three basic models, namely, the trend LSTM, the seasonal LSTM, and the original LSTM, are trained; and then, the ridge regression model is trained to fuse the predicted values of the three basic models to obtain the final predicted values. Finally, the method of this paper is applied to the heat load prediction of eight groups in a large mountain hotel park, and the root mean square error (RMSE) and mean absolute error (MAE) are used as the evaluation indexes. The experimental results show that the average RMSE and average MAE of the prediction results of the proposed method in this paper are minimized on the eight groups.

Improving prediction accuracy of CSM-CERES-Wheat model for water and nitrogen response using a modified Penman-Monteith equation in a semi-arid region

Context or problemThe implementation of crop models for water and nitrogen management are important in preserving water resources and reducing groundwater pollution. The Decision Support System for Agrotechnology Transfer (DSSAT) is a collection of field-scale process-based crop models that simulate crop growth and yield in response to water and nitrogen management as well as many other factors. Objective or research questionThe objective of this study was to improve the simulation of irrigation and nitrogen fertilizer effects on growth and yield of winter wheat using locally modified Penman-Monteith (PM) equation. Specifically, we focused on improving the estimation of ETo using a locally modified Penman-Monteith (PM) equation in the CSM-CERES-Wheat model. MethodsField measured data from a two-year experiment (2009–2010 and 2010–2011) in a semi-arid region were used to calibrate and evaluate the CSM-CERES-Wheat model in a region with strong advective condition. In this experiment, the irrigation treatments were 1.2, 1.0, 0.8, 0.5 times by full irrigation requirements and rain-fed, and nitrogen treatments were 0, 46, 92, and 136 kg N ha−1. ResultsUsing the CSM-CERES-Wheat model with the unmodified PM equation resulted in some growth variables such as: aboveground dry matter, grain yield, and grain nitrogen uptake with high normalized root mean square error (NRMSE) of 0.51, 0.35, and 0.32, respectively. However, using the modified PM equation for ETo module under semi-arid condition improved the soil water content estimation and the model outputs, such as the above ground dry matter, grain yield and grain nitrogen uptake with NRMSE of 0.2, 0.15 and 0.16, respectively, that are accurate. This modification provided more accurate estimations, particularly in our experimental site in semi-aid region with strong advective conditions. ConclusionsThe findings of this study indicated that the locally modified PM equation in the CSM-CERES-Wheat model enhanced the accuracy of estimating irrigation and nitrogen fertilizer on the growth and yield of winter wheat. By improving the estimation of ETo value, the model's performance in simulating crop responses to water and nitrogen management was significantly enhanced. Implications or significanceThe utilization of the accurately estimated ETo value of the modified PM equation has significant importance in the practical application of the CSM-CERES-Wheat model. Furthermore, this modification greatly enhances the accuracy of model outputs, particularly at our experimental site located in a semi-arid region with robust advective conditions.

A novel hierarchical approach to insight to spectral characteristics in surface water of karst wetlands and estimate its non-optically active parameters using field hyperspectral data

Wetlands cover only around 6 % of the Earth's land surface, and are recognized as one of the three major ecosystems, alongside forests and oceans. The ecological structure and function of karst wetlands are unique due to the influence of geologic structure. At present, the unclear spectral morphology of surface water in karst wetlands poses a significant challenge in remote sensing estimation of non-optically active water quality parameters (NAWQPs). This study proposed a novel multi-scale spectral morphology feature extraction (MSFE) method to insight to spectral characteristics in surface water of karst wetlands, and further screen the sensitive features of NAWQPs. Then we constructed three remote sensing inversion strategies for NAWQPs (TN, TP, NH3_N, DO), including direct estimation, indirect estimation, and auxiliary estimation. Finally, we constructed a novel pH-based hierarchical analysis framework (pH_HA) to thoroughly explore the influence of alkalinity-biased characteristics of karst water on the spectral domain of NAWQPs and its estimation accuracy using in-situ hyperspectral data, respectively. We found that the spectral characteristics of karst waters at the first reflectance peak (580 nm) differed significantly from other water body types. The MSFE successfully captured the sensitive spectral domains for NAWQPs, and focused on between 500 and 600 nm and 900–960 nm. The sensitive features captured by MSFE improved estimation accuracy of NAWQPs (R2 >0.9). Direct estimation presented more stable performance compared to the auxiliary estimation (average RMSE of 0.366 mg/L), and the auxiliary estimation model further improved the retrieval accuracy of TN compared to direct estimation model (R2 increasing from 0.43 to 0.56). The novel hierarchical framework clearly revealed the notable changes in the sensitive spectral domains of NAWQPs under different pH values, and enabled more precise determination of spectral subdomains of NAWQPs, and identified the optimal spectral features. The pH_HA framework effectively improved the estimation accuracy of NAWQPs (R2 increased from 0.514 to over 0.9), and the estimation accuracies (R2) of four NAWQPs were all more than 0.9 when the pH value was over 8.5. Our works provide an effective approach for monitoring water quality in karst wetlands.

Walk-Through Corrosion Assessment of Slurry Pipeline Using Machine Learning

The study of pipeline corrosion is crucial to prevent economic losses, environmental degradation, and worker safety. In this study, several machine learning methods such as recursive feature elimination (RFE), principal component analysis (PCA), gradient boosting method (GBM), support vector machine (SVM), random forest (RF), K-nearest neighbors (KNN), and multilayer perceptron (MLP) were used to estimate the thickness loss of a slurry pipeline subjected to erosion corrosion. These different machine learning models were applied to the raw data (the set of variables), to the variables selected by RFE, and to the variables selected by PCA (principal components), and a comparative analysis was carried out to find out the influence of the selection and transformation of the data on the performance of the models. The results show that the models perform better on the variables selected by RFE and that the best models are RF, SVM, and GBM with an average RMSE of 0.017. By modifying the hyperparameters, the SVM model becomes the best model with an RMSE of 0.011 and an R-squared of 0.83.