Mean Absolute Percentage Error Research Articles

The Impact of Drying Techniques on Stabilizing Microencapsulated Astaxanthin From Shrimp Shells: A Comparative Study of Spray Drying Versus Freeze Drying

ABSTRACTThe research aimed to study how different drying methods (spray and freeze drying) affect the release kinetics of microencapsulated astaxanthin in various environmental conditions. Shrimp shell extract containing astaxanthin was encapsulated using different wall components (maltodextrin with different dextrose equivalents and modified starch) via a simplex lattice mixture design. The encapsulated extract was then subjected to storage at different temperatures (25°C ± 2°C and 2°C ± 4°C) and humidity conditions (52% ± 2% and 75% ± 2%), as well as exposure to UV light (four 15 W lamps, 254 nm, for 10 h). The release kinetics of astaxanthin were analyzed using various models (page, Newton Korsmeyer–Peppas model, Modified Henderson and Pabis, Diffusion approach, and Two‐term exponential). The evaluation results of correlation coefficient (R2), root mean square deviation (RMSE), and mean absolute percentage error (MAPE) values of different models showed that the astaxanthin degradation followed a two‐term exponential kinetics in both types of microcapsules. Astaxanthin degradation increased with higher temperatures, humidity, and UV light exposure. However, microcapsules with equal wall compound ratios exhibited better preservation of astaxanthin. The study also emphasized the significance of optimizing storage conditions and wall materials for microencapsulated astaxanthin, as well as the utility of the two‐term exponential model in enhancing stability and shelf life.

Prediction of building cooling load: an innovative design for GA-SVM and IG-SVM system for the estimation of hourly precision and computational fluctuation range

ABSTRACT Anticipating the loads that buildings will bear is a crucial strategy in the pursuit of energy efficiency and the reduction of emissions. In this paper, we introduce an interactive and comprehensive joint prediction model, specifically crafted to elevate the precision of forecasts related to building loads. The resultant Genetic Algorithm-Support Vector Machine (GA-SVM) prediction model is put into action to provide hourly predictions of cooling loads for buildings. In scenarios where the prediction of building cooling load (BCL) fluctuations becomes imperative, especially in the face of extreme weather conditions, the information granulation (IG) method is employed. The findings of this validation process unveil significant improvements. The coefficient of variation of root mean square error (CV-RMSE) and mean absolute percentage error (MAPE) for the GA-SVM model are reduced by 58.85% and 68.04%, respectively, when compared to the conventional SVM model. Furthermore, in a comparative analysis against three widely utilized prediction models, the SVM model demonstrates a reduction in CV-RMSE and MAPE ranging from 2.04% to 68.04%. The application of the joint prediction model showcases impressive results, with R 2 values ranging from 97.27% to 99.68%, MAPE ranging from 2.59% to 2.84%, and CV-RMSE ranging from only 0.0249 to 0.0319.

ANN Prediction of Mode I Fracture Energy in Epoxy Adhesive Joints: Adherend, Adhesive, and Strain Rate Effects

In this investigation, the mode I fracture behavior of an epoxy adhesive joint has been investigated to probe the combined effects of adherend Young's modulus and thickness, strain rate, and adhesive length. Experimental fracture testing was conducted on double-cantilever beam (DCB) specimens under three various strain rate thresholds, namely, quasi-static (0.0005 s-1), low (0.015 s-1), and intermediate (0.5 s-1). The specimens were constructed with varying adherend materials (i.e., copper, aluminum), thickness (12-20 mm), and adhesive lengths (25-75 mm). The fracture load was measured in each experiment, and the critical strain energy release rate, JCi, was calculated via a finite element analysis (FEA) in each case. Results with a 95% confidence interval showed adherend thickness had a negligible effect within the investigated range (p-value = 0.462). JCi was significantly impacted by adhesive length (p-value = 0.009) and two other investigated parameters (p-values = 0.000). The development of artificial neural networks (ANNs) with Levenberg-Marquardt (LM) led to the prediction of JCi based on the examined parameters. With a mean absolute percentage error (MAPE) of 8.9% and a mean squared error (MSE) of 683 J2/m4 on unseen test data, the ANN model built in this study was able to predict the JCi in epoxy adhesive joints, indicating its potential to produce an accurate prediction for JCi in epoxy adhesive joints. This work offers insights for precise fracture behavior prediction under mode I stress conditions as well as some helpful information that can be utilized to optimize adhesive joint design.

Multi-relation spatiotemporal graph residual network model with multi-level feature attention: A novel approach for landslide displacement prediction

Accurate prediction of landslide displacement is crucial for effective early warning of landslide disasters. While most existing prediction methods focus on time-series forecasting for individual monitoring points, there is limited research on the spatiotemporal characteristics of landslide deformation. This paper proposes a novel Multi-Relation Spatiotemporal Graph Residual Network with Multi-Level Feature Attention (MFA-MRSTGRN) that effectively improves the prediction performance of landslide displacement through spatiotemporal fusion. This model integrates internal seepage factors as data feature enhancements with external triggering factors, allowing for accurate capture of the complex spatiotemporal characteristics of landslide displacement and the construction of a multi-source heterogeneous dataset. The MFA-MRSTGRN model incorporates dynamic graph theory and four key modules: multi-level feature attention, temporal-residual decomposition, spatial multi-relational graph convolution, and spatiotemporal fusion prediction. This comprehensive approach enables the efficient analyses of multi-source heterogeneous datasets, facilitating adaptive exploration of the evolving multi-relational, multi-dimensional spatiotemporal complexities in landslides. When applying this model to predict the displacement of the Liangshuijing landslide, we demonstrate that the MFA-MRSTGRN model surpasses traditional models, such as random forest (RF), long short-term memory (LSTM), and spatial temporal graph convolutional networks (ST-GCN) models in terms of various evaluation metrics including mean absolute error (MAE = 1.27 mm), root mean square error (RMSE = 1.49 mm), mean absolute percentage error (MAPE = 0.026), and R-squared (R2 = 0.88). Furthermore, feature ablation experiments indicate that incorporating internal seepage factors improves the predictive performance of landslide displacement models. This research provides an advanced and reliable method for landslide displacement prediction.

Communication—Prediction of Area Specific Resistance of Solid Oxide Cell Stacks from Electrochemical Impedance Spectra

A support vector regression model was trained to predict the area specific resistance of solid oxide cell stacks from electrochemical impedance spectroscopy measurements. In particular, the minimum necessary frequency range required for accurate predictions was investigated. Therefore, 711 pairs of DC polarization and electrochemical impedance spectroscopy measurements, conducted at the same stage of stack degradation and under similar measurement conditions, were evaluated. Prediction accuracies of less than 8% mean absolute percentage error and coefficients of determination greater than 93% were achieved. The 101–102 Hz range is sufficient for accurate predictions, allowing an efficient and non-destructive etimation of DC polarization measurements.

A spatio-temporal forecasting method of fracture distribution using dynamically exposed rock images in tunnel: Methodology and application

Forecasting the geometric characteristics of rock fractures in unexcavated tunnel sections is crucial for construction arrangements. This paper presents a method for predicting fracture distribution near the tunnel face from a spatiotemporal perspective. We Innovatively analogize the tunnel excavation mileage as the time series data. Using fractal geometry and geostatistics, we extract geometric features of peripheral rock fractures and establish a spatiotemporal dataset for dynamic prediction. The SCINet model, a spatial-aware recursive neural network, is employed for deterministic-stochastic dynamic prediction. Numerical simulations are conducted to validate the predictive accuracy of the model for various distributions, including trace length, dip angle, and density. Notably, the Mean Absolute Percentage Error (MAPE) for trace length prediction is remarkably low at 5.49 %, and the Kullback-Leibler (KL) divergence is as low as 4.81 %. The method is further applied to an underground oil storage cavern, China, revealing structural surface information. By incorporating continuously exposed fracture data, prediction accuracy is progressively improved, demonstrating the method's potential as a predictive tool for enhancing tunnel construction safety and efficiency.

Hybrid Photovoltaic and Gravity Energy Storage Integration for Smart Homes with Grid-Connected Management

This paper introduces a dynamic Smart Home Energy Management System (SHEMS) integrating a hybrid photovoltaic (PV) and gravity energy storage (GES) system aimed at minimizing environmental impacts and household energy consumption. The novel SHEMS features a one-week dynamic forecasting model that adapts to variable electricity prices, smart appliance schedules, solar output, and energy storage states. These results demonstrate that the system not only reduces household energy usage but also cuts electricity bills significantly, supplying power autonomously for up to 8.5 hours daily. By leveraging real-time data from the Dark Sky API on cloud cover and temperature, this model accurately predicts solar radiation and PV generation, aligning it with both grid and residential demands. The forecasting accuracy was assessed using Root Mean Square Error (RMSE) and Mean Absolute Percentage Error (MAPE), which improved to 12.55% and 4.91% respectively, from initial values of 22.46% for RMSE and 11.78% for MAPE. These advancements enhance grid stability and optimize energy storage during peak periods, reducing dependence on fossil fuels. The integration of innovative renewable energy technologies and sophisticated forecast modeling significantly boosts the system's efficiency, promoting the sustainable use of energy resources in line with environmental and economic goals.

GMOPNet: A GAN-MLP two-stage network for optical properties measurement of kiwifruit and peaches with spatial frequency domain imaging

Spatial frequency domain imaging (SFDI) is an imaging technique using spatially modulated illumination for measurement of optical properties. Conventional SFDI methods require capturing at least six images, making it time-consuming. This study presents a Generative Adversarial Network-Multi-Layer Perceptron (GAN-MLP) two-stage network (GMOPNet) for extracting high-precision optical properties of kiwifruit and peaches from a single SFDI image, enabling real-time continuous wide-band SFDI. The GMOPNet we proposed leverages the GAN to predict diffuse reflectance, followed by the MLP with Monte Carlo prior knowledge to predict optical properties. Our method achieves mean absolute percentage errors (MAPE) of 5.91 % for the absorption coefficient (μa) and 5.23 % for the reduced scattering coefficient (μs′), reducing acquisition and processing time significantly, with single inference taking 31.13 ms. The MAPE of the μa and the μs′ were 6.73 % and 6.34 % for kiwifruit and 5.80 % and 6.65 % for peaches, respectively.

Interpretable Multi-Horizon Time Series Forecasting of Cryptocurrencies by Leverage Temporal Fusion Transformer

This research delves into the obstacles and difficulties associated with predicting cryptocurrency movements in the volatile global financial market. This study develops and evaluates an advanced Deep Learning-Enhanced Temporal Fusion Transformer (ADE-TFT) model to estimate Bitcoin values more accurately. This research employs cutting-edge artificial intelligence (AI) and machine learning (ML) techniques to comprehensively examine various aspects of cryptocurrency forecasting, including geopolitical implications, market sentiment analysis, and pattern detection in transactional datasets. The study demonstrates that the ADE-TFT model outperforms its lower-layer counterparts in terms of forecasting accuracy, with reduced Mean Absolute Percentage Error (MAPE), Mean Squared Error (MSE), and Root Mean Square Error (RMSE) values, particularly when using a higher hidden layer configuration (h=8). The study emphasizes the importance of experimenting with different normalization strategies and utilizing various market-related data to enhance the model's performance. The results suggest that improving forecasting accuracy may require addressing these limitations and incorporating additional factors, such as market sentiment. By providing investors with more precise market predictions, the techniques and information presented in this research have the potential to significantly increase investor power in an unpredictable digital currency market, enabling wise investment choices.

Time series-based machine learning for forecasting multivariate water quality in full-scale drinking water treatment with various reagent dosages

Accurately predicting drinking water quality is critical for intelligent water supply management and for maintaining the stability and efficiency of water treatment processes. This study presents an optimized time series machine learning approach for accurately predicting multivariate drinking water quality, explicitly considering the time-dependent effects of reagent dosing. By leveraging data from a full-scale treatment plant, we constructed feature-engineered time series datasets incorporating influent water quality parameters, reagent dosages and effluent water characteristics. Seven predictive models, including both traditional machine learning (ML) and deep learning (DL) models were developed and rigorously evaluated against a naive mean baseline model. Our results demonstrate that traditional ML models, enhanced with time feature engineering, rivaled the performance of both widely used DL models and the naive mean baseline model. Specifically, an XGBoost model achieved superior prediction accuracy in dynamically forecasting four water quality characteristics at a 12-hour time lag step, outperforming the naive baseline model by 3-4% in terms of Mean Absolute Percentage Error (MAPE). This finding underscores the importance of incorporating a 12-hour interval to effectively capture the delayed impact of reagent dosing on water quality prediction. Furthermore, SHAP model interpretability analysis provided valuable insights into the XGBoost model's decision-making process, revealing its strong data-driven foundation aligned with established water treatment principles. This research highlights the significant potential of optimized machine learning techniques for enhancing water purification processes and enabling more informed, data-driven decision-making in the water supply industry.

DeepPipe: A Multi-Stage Knowledge-Enhanced Physics-Informed Neural Network for Hydraulic Transient Simulation of Multi-Product Pipeline

In the chemical pipelining industry, owing to the high-pressure transportation process, an accurate hydraulic transient simulation tool plays a central role in preventing the slack line flow and overpressure from causing pipeline operation treacherous. Nevertheless, the current model-driven method often faces challenges in balancing computational efficiency with accuracy, and the existing data-driven models struggle to produce explainable results from the physics perspectives since insufficient theoretical principles are incorporated into the model training. Additionally, the existing physics-informed learning architecture fails to achieve a gradient-balanced training, resulting from the significant magnitude difference in outputs and multiple loss terms. Consequently, a Multi-Stage Knowledge-Enhanced Physics-Informed Neural Network (MS-KE-PINN) is proposed for the hydraulic transient simulation of multi-product pipelines. To enforce the neural network producing simulation results with high consistency to physical laws, the governing equations, boundary, and initial condition are incorporated into the training process for an efficient mesh-free simulation. Then, considering that the significant magnitude difference between outputs can easily lead to deficient performance in the gradient descent, the magnitude conversion on the outputs and the equivalent conversion of the governing equations are implemented to enhance the training effect of the neural network. Subsequently, to tackle the imbalanced gradient of multiple loss terms with fixed weights, a multi-stage hierarchical training strategy is designed to improve the approximation capacity of the neural network. Numerical simulation cases demonstrate a better approximation function of the proposed model than the state-of-art models, while the mean absolute percentage errors yielded by MS-KE-PINN are reduced by 77.4 %, 88.7 %, and 87.8 % in three simulation operation conditions for pressure prediction. Furthermore, experimental investigations from a real-world multi-product pipeline suggest that the proposed model can still draw accurate simulation results even under complex and dynamic hydraulic transient scenarios in practice, with root mean squared errors reduced by 94.8 % and 80 % than that of the physics-informed neural network. To this end, the proposed model can conduct accurate and effective hydraulic transient analysis, thus ensuring the safe operation of the pipeline.

Predicting tensile and fracture parameters in polypropylene-based nanocomposites using machine learning with sensitivity analysis and feature impact evaluation

This study examines the efficacy of decision tree and AdaBoost algorithms in predicting mechanical and fracture parameters of polypropylene nanocomposites toughened with ethylene-based and propylene-based thermoplastic elastomers and reinforced with fumed silica and halloysite nanotube nanoparticles. The essential work of the fracture approach was utilized to study the fracture parameters, including elastic and plastic works of the blended polymer nanocomposites. The data were divided into 80 % for training and 20 % for testing. AdaBoost consistently achieved superior performance compared to the decision tree model in all variables throughout both the training and testing stages. During the testing phase, the AdaBoost model obtained R2 values of 0.90 for Young's modulus, 0.93 for elongation at break, 0.87 for tensile strength, 0.86 for plastic work, and 0.60 for elastic work. Also, the mean absolute percentage error for the AdaBoost model during the test phase was 3.10 % for Young's modulus, 3.25 % for tensile strength, 10.34 % for elastic work, 13.55 % for plastic work, and 24.78 % for elongation at break. Furthermore, a sensitivity analysis examining the effects of various features such as TPO type, nanoparticles, and nanoparticle type on mechanical properties reveals that TPO has the most significant overall influence. The results also include an analysis of the impact of the key features on each mechanical property based on the sensitivity analysis.

Potential of Near-Infrared Optical Techniques for Non-invasive Blood Glucose Measurement: A Pilot Study

BackgroundDiabetes mellitus manifested by escalated blood glucose, demands periodic or continuous glucose monitoring in the pursuit of improved control, optimal management, and appropriate medication. The existing state-of-the-art for blood glucose estimation are invasive methodologies necessitating the acquisition of blood sample through either a finger-stick or venipuncture. MethodsWe propose a dual Near-infrared wavelength integrated in a 3D printed enclosure as the optical prototype for non-invasive glucose estimation. We acquired data from both diabetic and healthy subjects with the developed system and subsequently validated the system. ResultsThe system demonstrated commendable clinical accuracy, as evidenced by the alignment of data pairs representing actual blood glucose levels and blood glucose levels predicted by our optical system within the A+B zones of the Parkes error grid and zones of no-risk and lower risk as defined by the surveillance error grid. We achieved compliant pairs percentage of 95.6%, which satisfies the accuracy requirements of the blood glucose monitoring surveillance study. The mean absolute percentage error attained with the proposed device (5.99%) was significant in predicting the blood glucose. ConclusionWe successfully deployed the NIR wavelengths functionality as the promising approach for glucose monitoring, offering new possibilities for improved medical interventions.

Photodamage prediction model and optimal lighting system for polychrome artworks in museum

When lighting polychrome artworks in museum, the light sources are required not only to provide high-quality visual performance, but also to minimize the photodamage. Currently, though there are mature methods to calculate visual performance, how to predict the degree of photodamage remains a challenge, which leads to a lack of light sources satisfying both visual and protection requirements. Our study conducted a series of 294 accelerated aging experiments spanning over 11 years for materials commonly used in Chinese polychrome artworks, using isoenergetic white-light D55 and 10 different wavelengths of narrow-band light as experimental light sources. The mathematical model predicting the photodamage degree caused by intensity (I), exposure time (t), the relative spectral power distribution of the light source (S(λ)), and the irradiated material's relative spectral responsivity (P(λ)) to polychrome artworks was established, and the mean absolute percentage error of the model, as verified by experiments, was 11.63 %. Combined with the existing visual performance calculation methods, an optimal lighting algorithm that simultaneously meets high-quality visual performance and minimum photodamage was developed, followed by developing a new Spectrally Tunable White LEDs (STWLEDs) lighting system with 10 LED chips. It is found that the STWLEDs can adjust and output the optimal spectra and recommended illuminance according to the characteristics of polychrome artworks, thus realizing lighting with minimum photodamage, meanwhile, satisfying high-quality visual performance. After testing, as well as satisfying the requirements of visual performance, the average photodamage is reduced by 40.95 % with STWLEDs compared with the traditional WLEDs.

The short-term wind power prediction based on a multi-layer stacked model of BO-CNN-BiGRU-SA

Wind power generation is influenced by various meteorological factors, exhibiting significant volatility and unpredictability. This variability presents considerable challenges for accurate wind power forecasting. In this study, we propose an innovative method for short-term wind power prediction that integrates a Bayesian-optimized Convolutional Neural Network (CNN), Bidirectional Gated Recurrent Units (BiGRU), and a Self-Attention Mechanism (SA) within a multi-layer architecture. Initially, we preprocess features using Pearson correlation analysis and input them into the CNN to investigate complex nonlinear spatial relationships among multiple feature variables and the current load. Subsequently, the BiGRU captures long-term dependencies from both forward and backward time sequences. Finally, we implement the Self-Attention Mechanism to weigh the features and generate the predicted wind power. We optimize the model's numerous hyperparameters utilizing a Bayesian algorithm. Through comparative ablation experiments with varying time segment lengths on wind farm datasets from four regions, our method significantly outperforms 11 models, including Long Short-Term Memory (LSTM), and surpasses several state-of-the-art (SOTA) prediction models, such as iTransformer, PatchTST, Non-stationary Transformers, TSMixer, and DLinear. The highest coefficient of determination (R²) achieved was 0.981, with the Symmetric Mean Absolute Percentage Error (SMAPE), Root Mean Square Error (RMSE), and Mean Absolute Error (MAE) decreasing by 11.22% to 62.04% compared to other models. The results demonstrate the predictive accuracy and generalization performance of our proposed model.

A Novel Approach for Forecasting Tourist Arrivals Using Web Search Data and Artificial Intelligence

The development of economic activity has been matched by growth in the tourism industry. According to information, the tourism industry is growing and both the number of domestic and international tourists visiting each year is expanding. Because of this quick expansion, there are now critical complications with the management of tourism, such as predicting the arrivals for travel, particularly when a lot of people are visiting appealing locations for particular periods. The proposed Artificial Fish Swarm Optimized Dynamic Gated Recurrent Unit (AFSO-DGRU) approach transforms the forecasting of demand for tourism by utilizing intelligence from swarms to improve forecasts and strategically adapting to fluctuating visitor structures. It ensures accurate and dynamic responses even during times of uncertainty when demand is high. The study used Google Trends to collect data from searches on the web and examine trends in tourist’s interest and demand for travel. By combining innovative artificial intelligence (AI) algorithms with realtime online search data, this study presents a novel way to improve the accuracy of visitor arrival predictions. The proposed method performs better than the existing methods to utilize the parameters such as mean absolute deviation called MAE (42.01), mean square error denoted by MSE (3059.85), mean absolute percentage error defined MAPE (1.34), and RMSPE or root mean square percentage error (1.43). This research utilizes web search data and AI to improve the accuracy of forecasting tourist arrivals, offering valuable insights for understanding tourism trends.

A novel characterization methodology for vapor-injected compressors: A comparative analysis with existing black-box models

In regions characterized by high temperature gradients, vapor compression systems often necessitate operation at very high pressure ratios resulting in a reduction in system capacity. Economized vapor injection compressors are used to avoid these issues, yet a precise predictive map for various compressor technologies with minimal data and relatively better performance remains unclear. This paper establishes a black-box compressor model to accurately predict compressor power, injection mass ratio, and evaporator mass flow rate in compressors with a single vapor injection port. This model is compared against three legacy models from literature and the ANN model, for reference. All five models are evaluated based on their ability to predict the aforementioned metrics. The proposed black-box model can predict the relevant metrics all within 5 % Mean Absolute Percentage Error (MAPE). Additionally, a refrigerant sensitivity analysis is performed with the black-box models. The model is trained using data from R410A and then used the coefficients to predict the performance of the same compressor when using R454B, and vice versa. It can estimate the evaporator mass flow rate with an accuracy within 3 %, the power within 2 %, and the injection mass ratio with MAPE less than 3 %.

Comparison of Inverse Distance Weighted and Thin Plate Spline Interpolation Methods in Projecting the Strength of the West Sumatra Earthquake

The Indonesian archipelago is situated in a highly active geological zone, making it prone to frequent earthquakes. West Sumatra, located on the west coast of central Sumatra, comprises lowland coastal areas and volcanic plateaus formed by the Barisan Mountains, covering a land area of 42,297.30 km² (2.17% of Indonesia's territory). This research aims to determine which interpolation method—Inverse Distance Weighted (IDW) and Thin Plate Spline (TPS)—provides more accurate predictions of earthquake strength in West Sumatra. The dataset consists of 229 earthquake events, divided into 90% for training (206 points) and 10% for testing (23 points). The training data was further subdivided into 80% training data 2 (164 points) and 20% validation data (42 points). The interpolation processes using the IDW and TPS methods were repeated 100 times, with the training 2 and validation data randomly shuffled in each iteration. Visualization of the interpolation results indicated that the earthquake magnitudes ranged from 2.0 to 4.5. Although the Mean Absolute Percentage Error (MAPE) values for the TPS method on the test and validation datasets were 16.42 and 14.29, respectively—slightly lower than the MAPE values for the IDW method—the t-test results showed no statistically significant difference between the two methods. Statistically, there is no significant difference between IDW and TPS in terms of predictive accuracy. However, researchers prefer the IDW method due to its computational efficiency and simplicity. Therefore, IDW is considered the most suitable method for analyzing earthquake strength in the West Sumatra region

Forecasting Daily Radiotherapy Patient Volumes in a Tertiary Hospital Using Autoregressive Integrated Moving Average (ARIMA) Models.

The purpose is to predict the volume of patients treated daily with radiotherapy using the autoregressive integrated moving average (ARIMA) model. In this retrospective study, data from the billing records detailing daily radiotherapy treatment sessions were extracted from the Hospital Information System and analyzed. The study included all patients treated from January 2004 to December 2022. The analysis was divided into two parts: First, the data were summarized using descriptive statistics. Second, time series forecasting with the implementation of an ARIMA model for estimating patient volumes. For the ARIMA modeling process, the Akaike Information Criterion (AIC) was used for classical model optimization. The Mean Absolute Percentage Error (MAPE) was used for evaluating between different models. Residual analysis was performed in each model using the Ljung-Box test, Jarque-Bera test, and heteroskedasticity test to identify autocorrelation, normal distribution, and variances that could undermine the reliability of the model. A total of 895,808 radiotherapy sessions were included in the study. The median number of radiotherapy sessions per day was 181 (150, 205). A clear transition to more modern radiotherapy equipment, particularly the Truebeam accelerator, was observed, indicating a growing dependency on advanced techniques such as volumetric-modulated arc therapy (VMAT), stereotactic body radiation therapy (SBRT), and stereotactic radiosurgery (SRS). The best ARIMA model predicted an increase in demand, projecting an average daily patient volume of 279.40 by 2030. The study highlights the need for advanced forecasting methodologies in healthcare resource planning and emphasizes the importance of considering environmental and external factors for effective and accurate resource allocation strategies.

Characterizing Chromophoric Dissolved Organic Matter Spatio-Temporal Variability in North Andean Patagonian Lakes Using Remote Sensing Information and Environmental Analysis

Chromophoric dissolved organic matter (CDOM) is crucial in aquatic ecosystems, influencing light penetration and biogeochemical processes. This study investigates the CDOM variability in seven oligotrophic lakes of North Andean Patagonia using Landsat 8 imagery. An empirical band ratio model was calibrated and validated for the estimation of CDOM concentrations in surface lake water as the absorption coefficient at 440 nm (acdom440, m−1). Of the five atmospheric corrections evaluated, the QUAC (Quick Atmospheric Correction) method demonstrated the highest accuracy for the remote estimation of CDOM. The application of separate models for deep and shallow lakes yielded superior results compared to a combined model, with R2 values of 0.76 and 0.82 and mean absolute percentage errors (MAPEs) of 14% and 22% for deep and shallow lakes, respectively. The spatio-temporal variability of CDOM was characterized over a five-year period using satellite-derived acdom440 values. CDOM concentrations varied widely, with very low values in deep lakes and moderate values in shallow lakes. Additionally, significant seasonal fluctuations were evident. Lower CDOM concentrations were observed during the summer to early autumn period, while higher concentrations were observed in the winter to spring period. A gradient boosting regression tree analysis revealed that inter-lake differences were primarily influenced by the lake perimeter to lake area ratio, mean lake depth, and watershed area to lake volume ratio. However, seasonal CDOM variation was largely influenced by Lake Nahuel Huapi water storage (a proxy for water level variability at a regional scale), followed by precipitation, air temperature, and wind. This research presents a robust method for estimating low to moderate CDOM concentrations, improving environmental monitoring of North Andean Patagonian Lake ecosystems. The results deepen the understanding of CDOM dynamics in low-impact lakes and its main environmental drivers, enhance the ability to estimate lacustrine carbon stocks on a regional scale, and help to predict the effects of climate change on this important variable.