Individual Machine Learning Models Research Articles

Enhancing Crop Yield Predictions with PEnsemble 4: IoT and ML-Driven for Precision Agriculture

This research introduces the PEnsemble 4 model, a weighted ensemble prediction model that integrates multiple individual machine learning models to achieve accurate maize yield forecasting. The model incorporates unmanned aerial vehicle (UAV) imagery and Internet of Things (IoT)-based environmental data, providing a comprehensive and data-driven approach to yield prediction in maize cultivation. Considering the projected growth in global maize demand and the vulnerability of maize crops to weather conditions, improved prediction capabilities are of paramount importance. The PEnsemble 4 model addresses this need by leveraging comprehensive datasets encompassing soil attributes, nutrient composition, weather conditions, and UAV-captured vegetation imagery. By employing a combination of Huber and M estimates, the model effectively analyzes temporal patterns in vegetation indices, in particular CIre and NDRE, which serve as reliable indicators of canopy density and plant height. Notably, the PEnsemble 4 model demonstrates a remarkable accuracy rate of 91%. It advances the timeline for yield prediction from the conventional reproductive stage (R6) to the blister stage (R2), enabling earlier estimation and enhancing decision-making processes in farming operations. Moreover, the model extends its benefits beyond yield prediction, facilitating the detection of water and crop stress, as well as disease monitoring in broader agricultural contexts. By synergistically integrating IoT and machine learning technologies, the PEnsemble 4 model presents a novel and promising solution for maize yield prediction. Its application holds the potential to revolutionize crop management and protection, contributing to efficient and sustainable farming practices.

Unleashing the power of machine learning and remote sensing for robust seasonal drought monitoring: A stacking ensemble approach

Droughts cause significant economic losses in many regions around the world, highlighting a need to more accurately quantify implications of drought on production and water management. Remote sensing technologies and machine learning-based models offer promising solutions for timely and accurate regional drought monitoring, but the accuracy of such approaches is constrained by both the conceptual design and algorithms underpinning such approaches. We developed a machine learning stacking ensemble approach to overcome such limitations, drawing upon Precipitation Estimation from Remotely Sensed Information using Artificial Neural Network-Climate Data Record (PERSIANN-CDR), Moderate Resolution Imaging Spectroradiometer (MODIS) remote sensing products, and climate zoning data to estimate 3-month scale SPEI (Standardized Precipitation Evapotranspiration Index, SPEI-3) in 9 sub-regions of China. We compared 19 individual machine learning models and used the stacking approach to select the most robust meta-model for SPEI-3 prediction. We found that CatBoost Regressor (CBR), Extra Trees Regressor (ETR), Extreme Gradient Boosting (XGB), Light Gradient Boosting Machine (LGBM), and Random Forest (RF) were the top individual models for predicting drought, while using CBR as the stacked meta-model achieved the best performance. The R2 values for the stacking model with CBR as the meta-model were 0.9065 and 0.8218 in the eastern and western regions, respectively. We then employed the stacking model with CBR as the meta-model to generate seasonal drought maps based on SPEI-3 across different years and seasons. Our predicted SPEI-3 was compared with drought maps generated by the GPCC and ERA5 reanalysis datasets. We found a strong correlation between them with R2 exceeding 0.8 across multiple years and seasons, indicating our machine learning stacking ensemble approach had a good performance in monitoring seasonal drought conditions. The findings of our study establish a standardized protocol for predicting droughts in the agricultural sector, enabling its widespread application irrespective of variations in terrain and climate.

Comparative analysis of single and hybrid machine learning models for daily solar radiation

This study investigates the estimation of daily solar radiation (SR) through various machine learning (ML) models, including the k-nearest neighbor algorithm (KNN), support vector regression (SVR), and random forest (RF), both individually and in combination with the wavelet transform (WT). The assessment of these models is based on meteorological data spanning three decades (1981–2010) from the province of Kütahya in Türkiye. To address the inherent uncertainty in these data-driven models, the quantile regression method is employed for uncertainty analysis. Statistical metrics, such as mean absolute error (MAE), root mean square error (RMSE), coefficient of determination (R2), mean prediction interval (MPI), and prediction interval coverage probability (PICP), are utilized to evaluate the effectiveness and uncertainties of the models. The SVR and KNN models exhibit comparable performances concerning both predictive accuracy and uncertainty levels. However, hybrid models, such as KNN-WT, RF-WT, and SVR-WT, display enhanced accuracy compared to individual ML models, as indicated by statistical performance criteria. Notably, the SVR-WT model, incorporating inputs such as sunshine duration, air temperature, wind speed, and relative humidity, outperforms other models in terms of RMSE (2.174 MJ/m2), MAE (1.721 MJ/m2), R2 (0.923), MPI (28.55), and PICP (0.80) for the testing dataset. In conclusion, the integration of WT significantly improves the performance of ML models, providing valuable insights for the design and operation of solar energy systems, where precise daily SR estimation is critical for optimal and cost-efficient operation.

Urban waterlogging susceptibility assessment based on hybrid ensemble machine learning models: A case study in the metropolitan area in Beijing, China

Urban waterlogging has emerged as a significant problem worldwide, particularly in densely populated cities. Accurate assessment of waterlogging susceptibility at the city scale is crucial for mitigating the risks associated with waterlogging and optimizing municipal design accordingly. However, existing studies on urban waterlogging susceptibility assessment have primarily relied on individual machine learning models. It is worthwhile to explore whether hybrid ensemble models have the potential to enhance the predictive performance. This research presents two hybrid ensemble machine learning models, namely Stacking and Blending, for assessing urban waterlogging susceptibility in the metropolitan area of Beijing, China. The performances of these models are compared with those of the widely used individual machine learning models. Evaluation of all the models is based on metrics such as Accuracy rate and Area Under Curve (AUC) score. The results demonstrate that the Stacking and Blending models consistently outperform the traditional machine learning models, such as Random Forest, Logistic Regression, etc. Through susceptibility analysis and model interpretation with SHAP method, this paper obtains several key findings that low lying areas may not necessarily be areas with severe waterlogging; urban roads and densely populated areas are highly susceptible to becoming high-risk areas for waterlogging in the study area. This study not only highlights the effectiveness of the Stacking and Blending models for urban waterlogging susceptibility assessment but also provides valuable insights for waterlogging mitigation strategies in urban planning.

Forest Fire Risk Prediction Based on Stacking Ensemble Learning for Yunnan Province of China

Forest fire risk prediction is essential for building a forest fire defense system. Ensemble learning methods can avoid the problem of difficult model selection for disaster susceptibility prediction and can significantly improve modeling accuracy. This study introduces a stacking ensemble learning model for predicting forest fire risks in Yunnan Province by integrating various data types, such as meteorological, topographic, vegetation, and human activity factors. A total of 70,274 fire points and an equal number of randomly selected nonfire points were used to develop the model, with 70% of the data allocated for training and the remaining 30% for testing. The stacking model combined four diverse machine learning methods: random forest (RF), extreme gradient boosting (XGBoost), light gradient boosting machine (LightGBM), and multilayer perceptron (MLP). We evaluated the model’s predictive performance using metrics like accuracy, area under the characteristic curve (AUC), and fire density (FD). The results demonstrated that the stacking fusion model exhibited remarkable accuracy with an AUC of 0.970 on the test set, significantly surpassing the performance of individual machine learning models, which had AUC values ranging from 0.935 to 0.953. Furthermore, the stacking fusion model effectively captured the maximum fire density in extremely high susceptibility areas, demonstrating enhanced generalization capabilities.

Enhancing Network Intrusion Detection Using an Ensemble Voting Classifier for Internet of Things.

In the context of 6G technology, the Internet of Everything aims to create a vast network that connects both humans and devices across multiple dimensions. The integration of smart healthcare, agriculture, transportation, and homes is incredibly appealing, as it allows people to effortlessly control their environment through touch or voice commands. Consequently, with the increase in Internet connectivity, the security risk also rises. However, the future is centered on a six-fold increase in connectivity, necessitating the development of stronger security measures to handle the rapidly expanding concept of IoT-enabled metaverse connections. Various types of attacks, often orchestrated using botnets, pose a threat to the performance of IoT-enabled networks. Detecting anomalies within these networks is crucial for safeguarding applications from potentially disastrous consequences. The voting classifier is a machine learning (ML) model known for its effectiveness as it capitalizes on the strengths of individual ML models and has the potential to improve overall predictive performance. In this research, we proposed a novel classification technique based on the DRX approach that combines the advantages of the Decision tree, Random forest, and XGBoost algorithms. This ensemble voting classifier significantly enhances the accuracy and precision of network intrusion detection systems. Our experiments were conducted using the NSL-KDD, UNSW-NB15, and CIC-IDS2017 datasets. The findings of our study show that the DRX-based technique works better than the others. It achieved a higher accuracy of 99.88% on the NSL-KDD dataset, 99.93% on the UNSW-NB15 dataset, and 99.98% on the CIC-IDS2017 dataset, outperforming the other methods. Additionally, there is a notable reduction in the false positive rates to 0.003, 0.001, and 0.00012 for the NSL-KDD, UNSW-NB15, and CIC-IDS2017 datasets.

Predicting dissolved oxygen level using Young's double-slit experiment optimizer-based weighting model

Accurate prediction of the dissolved oxygen level (DOL) is important for enhancing environmental conditions and facilitating water resource management. However, the irregularity and volatility inherent in DOL pose significant challenges to achieving precise forecasts. A single model usually suffers from low prediction accuracy, narrow application range, and difficult data acquisition. This study proposes a new weighted model that avoids these problems, which could increase the prediction accuracy of the DOL. The weighting constructs of the proposed model (PWM) included eight neural networks and one statistical method and utilized Young's double-slit experimental optimizer as an intelligent weighting tool. To evaluate the effectiveness of PWM, simulations were conducted using real-world data acquired from the Tualatin River Basin in Oregon, United States. Empirical findings unequivocally demonstrated that PWM outperforms both the statistical model and the individual machine learning models, and has the lowest mean absolute percentage error among all the weighted models. Based on two real datasets, the PWM can averagely obtain the mean absolute percentage errors of 1.0216%, 1.4630%, and 1.7087% for one-, two-, and three-step predictions, respectively. This study shows that the PWM can effectively integrate the distinctive merits of deep learning methods, neural networks, and statistical models, thereby increasing forecasting accuracy and providing indispensable technical support for the sustainable development of regional water environments.

Improving PM2.5 prediction in New Delhi using a hybrid extreme learning machine coupled with snake optimization algorithm

Fine particulate matter (PM2.5) is a significant air pollutant that drives the most chronic health problems and premature mortality in big metropolitans such as Delhi. In such a context, accurate prediction of PM2.5 concentration is critical for raising public awareness, allowing sensitive populations to plan ahead, and providing governments with information for public health alerts. This study applies a novel hybridization of extreme learning machine (ELM) with a snake optimization algorithm called the ELM-SO model to forecast PM2.5 concentrations. The model has been developed on air quality inputs and meteorological parameters. Furthermore, the ELM-SO hybrid model is compared with individual machine learning models, such as Support Vector Regression (SVR), Random Forest (RF), Extreme Learning Machines (ELM), Gradient Boosting Regressor (GBR), XGBoost, and a deep learning model known as Long Short-Term Memory networks (LSTM), in forecasting PM2.5 concentrations. The study results suggested that ELM-SO exhibited the highest level of predictive performance among the five models, with a testing value of squared correlation coefficient (R2) of 0.928, and root mean square error of 30.325 µg/m3. The study's findings suggest that the ELM-SO technique is a valuable tool for accurately forecasting PM2.5 concentrations and could help advance the field of air quality forecasting. By developing state-of-the-art air pollution prediction models that incorporate ELM-SO, it may be possible to understand better and anticipate the effects of air pollution on human health and the environment.

Hybridization of Machine Learning Algorithms and an Empirical Regression Model for Predicting Debris-Flow-Endangered Areas

Accurate delineation of debris-flow-endangered areas (e.g., the maximum runout distance) is a necessary prerequisite for the debris-flow risk assessment and countermeasures design. Recently, machine-learning models have been proved to be an effective tool in predicting debris-flow parameters. However, existing machine-learning models are generally developed based on a very limited number of observation data, which may result in the predictive model overfitting or underfitting. How to develop a robust model for accurate forecasting of debris-flow-endangered areas still remains a difficult task. This paper proposes a hybrid method for predicting debris-flow hazard zone by integrating machine-learning algorithms and an empirical regression model. The proposed method takes the calculated maximum runout distance obtained from the empirical model as supplementary inputs to increase the amount of training data to construct hybrid machine-learning models. Three commonly used machine-learning models (i.e., multivariate adaptive regression splines (MARS), random forest (RF), and support vector machine (SVM)) are developed based on the training datasets of a specific debris basin. Then, these three machine-learning models are combined with an empirical relationship developed using the same training datasets to generate corresponding hybrid models. Finally, the performance metrics (i.e., coefficient of determination (R2), root-mean-square error (RMSE), and mean absolute error (MAE)) of the proposed hybrid models are comprehensively investigated and compared with the single predictive model (i.e., MARS, RF, SVM, and the empirical model) under fivefold cross-validation. The proposed method is illustrated using 134 channelized debris-flow events in Sichuan province, China. Results show that compared with the three individual machine-learning models, hybridization of machine-learning algorithms and the empirical model results in R2, RMSE, and MAE improved by 70.5%, 32.9%, and 41.1%, respectively. In contrast to the empirical model, the R2, RMSE, and MAE value of the proposed hybrid models are improved by 29.6%, 22.3%, and 32.5%, respectively. The proposed hybrid models generally perform better than the single machine-learning and the empirical model, providing a promising tool for accurate forecasting of a debris-flow-endangered area.

Estimation of soil moisture using environmental covariates and machine learning algorithms in Cathedral Peak Catchment, South Africa

AbstractSoil moisture (SM) is a fundamental constituent of the terrestrial environment and the hydrological cycle. Owing to its significant influence on catchment hydrological responses, it can be utilized as an indicator of floods and droughts to aid early warning systems. This study aimed to develop a field‐scale method to estimate SM using parametric and machine learning‐based methods to compare whether advanced artificial intelligence methods can give similar results as traditional methods. Considering this, monthly observed SM data (from the top 10 cm), environmental covariates, and remotely sensed data from March 2019 to July 2021 for the Cathedral Peak Research Catchments VI and IX in South Africa were obtained. From the 241 observations obtained across 12 sites, 160 (∼66%) were used for model training, while the remaining 81 (∼34%) were used for model testing. Employing 10‐fold cross‐validation, the individual machine learning models (viz., support vector machine [SVM], random forest (RF), k‐nearest neighbor, classification and regression trees [Rpart], and generalized linear model) displayed a satisfactory performance (R2 = 0.52–0.79; root mean square error = 3.79–5.80). In the validation phase, the RF model displayed a superior performance, followed by the SVM. Subsequent SM estimation using the hybrid model produced satisfactory results in training (R2 = 0.90) and testing (R2 = 0.45). The results obtained from this study can aid in predicting SM variations in catchments with limited monitoring. Furthermore, this model can be applied in drought monitoring, forecasting, and informing agricultural management practices.

Stock market analysis using ensemble learning

A stock market or share market is the combination of shoppers and sellers of shares. Prediction of the stock market is a method for calculating the future value of a company's stock. Stock market can be regarded as a specific records of data mining as well as machine learning problem. The daily changes within the stock depends on the profits and losses and many people think that stock market is irregular and uncertain. Based on daily changes we can predict some movements in the stock. In the previous years, researchers have used many machine learning models to know the development of the stock market to enhance the accuracy. However researchers have implemented these machine learning models independently and compared the results of models. It is observed that the group of models produce quite much less noisy compared to independent models. Our vision is to study investment strategies to predict and analyses the stocks using machine learning models such as Random Forest, K Nearest Neighbor, Gaussian Naive bayes and Decision Tree. And these models are ensemble by using different techniques such as Voting Classifier, Adaboost classifier and Bagging Classifier to estimate their accuracies. From the results it is observed that ensemble approach gives maximum accuracy compared to individual machine learning models.

Prediction of shear strength in UHPC beams using machine learning-based models and SHAP interpretation

To provide more accurate and reliable predictions of the shear strength of ultrahigh-performance concrete (UHPC) beams, in this study, the machine learning (ML) approaches were employed to develop the data-driven models, and the ML models were interpreted using the Shapley additive explanations (SHAP) method. It was found that the ensemble models, particularly CatBoost, outperform individual ML models and traditional empirical models. The geometric dimensions and shear span-to-depth ratio were the most influential features for predicting the shear strength of UHPC beams, followed by the parameters of reinforcement and material properties of the UHPC.

The Diagnosis of Malignant Pleural Effusion Using Tumor-Marker Combinations: A Cost-Effectiveness Analysis Based on a Stacking Model.

By incorporating the cost of multiple tumor-marker tests, this work aims to comprehensively evaluate the financial burden of patients and the accuracy of machine learning models in diagnosing malignant pleural effusion (MPE) using tumor-marker combinations. Carcinoembryonic antigen (CEA), carbohydrate antigen (CA)19-9, CA125, and CA15-3 were collected from pleural effusion (PE) and peripheral blood (PB) of 319 patients with pleural effusion. A stacked ensemble (stacking) model based on five machine learning models was utilized to evaluate the diagnostic accuracy of tumor markers. We evaluated the discriminatory accuracy of various tumor-marker combinations using the area under the curve (AUC), sensitivity, and specificity. To evaluate the cost-effectiveness of different tumor-marker combinations, a comprehensive score (C-score) with a tuning parameter w was proposed. In most scenarios, the stacking model outperformed the five individual machine learning models in terms of AUC. Among the eight tumor markers, the CEA in PE (PE.CEA) showed the best AUC of 0.902. Among all tumor-marker combinations, the PE.CA19-9 + PE.CA15-3 + PE.CEA + PB.CEA combination (C9 combination) achieved the highest AUC of 0.946. When w puts more weight on the cost, the highest C-score was achieved with the single PE.CEA marker. As w puts over 0.8 weight on AUC, the C-score favored diagnostic models with more expensive tumor-marker combinations. Specifically, when w was set to 0.99, the C9 combination achieved the best C-score. The stacking diagnostic model using PE.CEA is a relatively accurate and affordable choice in diagnosing MPE for patients without medical insurance or in a low economic level. The stacking model using the combination PE.CA19-9 + PE.CA15-3 + PE.CEA + PB.CEA is the most accurate diagnostic model and the best choice for patients without an economic burden. From a cost-effectiveness perspective, the stacking diagnostic model with PE.CA19-9 + PE.CA15-3 + PE.CEA combination is particularly recommended, as it gains the best trade-off between the low cost and high effectiveness.

Advancing ocean subsurface thermal structure estimation in the Pacific Ocean: A multi-model ensemble machine learning approach

Estimation of the ocean subsurface thermal structure (OSTS) is important for understanding thermodynamic processes and climate variability. In the present study, a novel multi-model ensemble machine learning (Ensemble-ML) model is developed to retrieve subsurface thermal structure in the Pacific Ocean by integrating sea surface data with Argo observations. The Ensemble-ML model integrates four individual machine learning models to enhance estimation accuracy and reliability. Our results exhibit good agreement between the satellite sea surface temperature (SST) and sea surface salinity (SSS) data and Argo observations, providing validation for the utilization of these datasets in the Ensemble-ML model. The Ensemble-ML model exhibits better performance compared to individual machine learning models, with an average root mean square error (RMSE) of 0.3273 °C and an average coefficient of determination (R²) of 0.9905. Notably, incorporating geographical information as input variables enhance model performance, emphasizing the importance of considering spatial context in OSTS estimation. The Ensemble-ML model accurately captures the spatial distribution of OSTS across depths and seasons in the Pacific Ocean, effectively reproducing critical temperature features while maintaining strong agreement with Argo observations. Nevertheless, its performance shows relative weakness within the thermocline layer and the equatorial Pacific region (spanning from 10°S to 10°N latitude), which are characterized by complex circulation systems. Despite these challenges, the Ensemble-ML model effectively reproduces the spatial distribution of OSTS of the Pacific Ocean. This indicates the potential of machine learning models, particularly ensemble models, for enhancing OSTS estimation in the Pacific Ocean and other regions, offering valuable insights for future research and applications in physical oceanography.

Online load-loss risk assessment based on stacking ensemble learning for power systems

Power systems faces significant uncertainty during operation owing to the increased integration of renewable energy into power grids and the expansion of the scale of power systems, these factors lead to higher load-loss risks; therefore, realization of a fast online load-loss risk assessment is crucial to ensuring the operational safety and reliability of power systems. This paper presents an online load-loss risk assessment method for power systems based on stacking ensemble learning. First, a traditional load-loss risk assessment method based on power flow analysis was constructed to generate risk samples. The label of the sample is load-loss risk assessment index and the features are multiple operational variables of the power system. And the recursive feature elimination using cross validation (RFECV) was adopted for feature selection. Second, four different machine learning models, including support vector regression (SVR), extremely randomized trees (ET), extreme gradient boosting (XGBoost) and elastic network (EN) were used to form a stacking ensemble learning model for sample training. Moreover, to further improve the model performance, the particle swarm optimization (PSO) algorithms was used for parameter optimization. Finally, based on this model, the online load-loss risk assessment of a power system was realized. The application of the proposed method on IEEE test systems demonstrated that the proposed method was more accurate than methods based on individual machine learning models, from which the stacking was designed, while still maintaining a significant advantage in terms of runtime compared to the traditional risk assessment method.

Application of machine learning (individual vs stacking) models on MERRA-2 data to predict surface PM2.5 concentrations over India

The spatial coverage of PM2.5 monitoring is non-uniform across India due to the limited number of ground monitoring stations. Alternatively, Modern-Era Retrospective Analysis for Research and Applications, Version 2 (MERRA-2), is an atmospheric reanalysis data used for estimating PM2.5. MERRA-2 does not explicitly measure PM2.5 but rather follows an empirical model. MERRA-2 data were spatiotemporally collocated with ground observation for validation across India. Significant underestimation in MERRA-2 prediction of PM2.5 was observed over many monitoring stations ranging from −20 to 60 μg m−3. The utility of Machine Learning (ML) models to overcome this challenge was assessed. MERRA-2 aerosol and meteorological parameters were the input features used to train and test the individual ML models and compare them with the stacking technique. Initially, with 10% of randomly selected data, individual model performance was assessed to identify the best model. XGBoost (XGB) was the best model (r2 = 0.73) compared to Random Forest (RF) and LightGBM (LGBM). Stacking was then applied by keeping XGB as a meta-regressor. Stacked model results (r2 = 0.77) outperformed the best standalone estimate of XGB. Stacking technique was used to predict hourly and daily PM2.5 in different regions across India and each monitoring station. The eastern region exhibited the best hourly prediction (r2 = 0.80) and substantial reduction in Mean Bias (MB = −0.03 μg m−3), followed by the northern region (r2 = 0.63 and MB = −0.10 μg m−3), which showed better output due to the frequent observation of PM2.5 >100 μg m−3. Due to sparse data availability to train the ML models, the lowest performance was for the central region (r2 = 0.46 and MB = −0.60 μg m−3). Overall, India's PM2.5 prediction was good on an hourly basis compared to a daily basis using the ML stacking technique.

Thyroid Nodules Classification using Weighted Average Ensemble and D-CRITIC based TOPSIS Methods for Ultrasound Images.

Thyroid disorders are prevalent worldwide and impact many people. The abnormal growth of cells in the thyroid gland region is very common and even found in healthy people. These abnormal cells can be cancerous or non-cancerous, so early detection of this disease is the only solution for minimizing the death rate or maximizing a patient's survival rate. Traditional techniques to detect cancerous nodules are complex and time-consuming; hence, several imaging algorithms are used to detect the malignant status of thyroid nodules timely. This research aims to develop computer-aided diagnosis tools for malignant thyroid nodule detection using ultrasound images. This tool will be helpful for doctors and radiologists in the rapid detection of thyroid cancer at its early stages. The individual machine learning models are inferior to medical datasets because the size of medical image datasets is tiny, and there is a vast class imbalance problem. These problems lead to overfitting; hence, accuracy is very poor on the test dataset. This research proposes ensemble learning models that achieve higher accuracy than individual models. The objective is to design different ensemble models and then utilize benchmarking techniques to select the best model among all trained models. This research investigates four recently developed image transformer and mixer models for thyroid detection. The weighted average ensemble models are introduced, and model weights are optimized using the hunger games search (HGS) optimization algorithm. The recently developed distance correlation CRITIC (D-CRITIC) based TOPSIS method is utilized to rank the models. Based on the TOPSIS score, the best model for an 80:20 split is the gMLP+ViT model, which achieved an accuracy of 89.70%, whereas using a 70:30 data split, the gMLP+FNet+Mixer-MLP has achieved the highest accuracy of 82.18% on the publicly available thyroid dataset. This study shows that the proposed ensemble models have better thyroid detection capabilities than individual base models for the imbalanced thyroid ultrasound dataset.

Scrutinization of land subsidence rate using a supportive predictive model: Incorporating radar interferometry and ensemble soft-computing

Land subsidence is a huge challenge that land and water resource managers are still facing. Radar datasets revolutionize the way and give us the ability to provide information about it, thanks to their low cost. But identifying the most important drivers need for the modeling process. Machine learning methods are especially top of mind amid the prediction studies of natural hazards and hit new heights over the last couple of years. Hence, putting an efficient approach like integrated radar-and-ensemble-based method into practice for land subsidence rate simulation is not available yet which is the main aim of this research. In this study, the number of 52 pairs of radar images were used to identify subsidence from 2014 to 2019. Then, using the simulated annealing (SA) algorithm the key variables affecting land subsidence were identified among the topographical parameters, aquifer information, land use, hydroclimatic variables, and geological and soil factors. Afterward, three individual machine learning models (including Support Vector Machine, SVM; Gaussian Process, GP; Bayesian Additive Regression Tree, BART) along with three ensemble learning approaches were considered for land subsidence rate modeling. The results indicated that the subsidence varies between 0 and 59 cm in this period. Comparing the Radar results with the permanent geodynamic station exhibited a very strong correlation between the ground station and the radar images (R2 = 0.99, RMSE = 0.008). Parsing the input data by the SA indicated that key drivers are precipitation, elevation, percentage of fine-grained materials in the saturated zone, groundwater withdrawal, distance to road, groundwater decline, and aquifer thickness. The performance comparison indicated that ensemble models perform better than individual models, and among ensemble models, the nonlinear ensemble approach (i.e., BART model combination) provided better performance (RMSE = 0.061, RSR = 0.42, R2 = 0.83, PBIAS = 2.2). Also, the distribution shape of the probability density function in the non-linear ensemble model is much closer to the observations. Results indicated that the presence of significant fine-grained materials in unconsolidated aquifer systems can clarify the response of the aquifer system to groundwater decline, low recharge, and subsequent land subsidence. Therefore, the interaction between these factors can be very dangerous and intensify subsidence.

Machine learning-based prediction of in-ICU mortality in pneumonia patients

Conventional severity-of-illness scoring systems have shown suboptimal performance for predicting in-intensive care unit (ICU) mortality in patients with severe pneumonia. This study aimed to develop and validate machine learning (ML) models for mortality prediction in patients with severe pneumonia. This retrospective study evaluated patients admitted to the ICU for severe pneumonia between January 2016 and December 2021. The predictive performance was analyzed by comparing the area under the receiver operating characteristic curve (AU-ROC) of ML models to that of conventional severity-of-illness scoring systems. Three ML models were evaluated: (1) logistic regression with L2 regularization, (2) gradient-boosted decision tree (LightGBM), and (3) multilayer perceptron (MLP). Among the 816 pneumonia patients included, 223 (27.3%) patients died. All ML models significantly outperformed the Simplified Acute Physiology Score II (AU-ROC: 0.650 [0.584–0.716] vs 0.820 [0.771–0.869] for logistic regression vs 0.827 [0.777–0.876] for LightGBM 0.838 [0.791–0.884] for MLP; P < 0.001). In the analysis for NRI, the LightGBM and MLP models showed superior reclassification compared with the logistic regression model in predicting in-ICU mortality in all length of stay in the ICU subgroups; all age subgroups; all subgroups with any APACHE II score, PaO2/FiO2 ratio < 200; all subgroups with or without history of respiratory disease; with or without history of CVA or dementia; treatment with mechanical ventilation, and use of inotropic agents. In conclusion, the ML models have excellent performance in predicting in-ICU mortality in patients with severe pneumonia. Moreover, this study highlights the potential advantages of selecting individual ML models for predicting in-ICU mortality in different subgroups.

A Reinforcement Learning Approach for Ensemble Machine Learning Models in Peak Electricity Forecasting

Electricity peak load forecasting plays an important role in electricity generation capacity planning to ensure reliable power supplies. To achieve high forecast accuracy, multiple machine learning models have been implemented to forecast the monthly peak load in Thailand over the past few years, yielding promising results. One approach to further improve forecast accuracy is to effectively select the most accurate forecast value for each period from among the forecast values generated by these models. This article presents a novel reinforcement learning approach using the double deep Q-network (Double DQN), which acts as a model selector from a pool of available models. The monthly electricity peak load data of Thailand from 2004 to 2017 are used to demonstrate the effectiveness of the proposed method. A hyperparameter tuning methodology using a fractional factorial design is implemented to significantly reduce the number of required experimental runs. The results indicate that the proposed selection model using Double DQN outperforms all tested individual machine learning models in terms of mean square error.