LightGBM Research Articles

Physics guided data-driven model to estimate minimum miscibility pressure (MMP) for hydrocarbon gases

Local or regional availability of hydrocarbon gases, and being associated with the native fluids makes them a reliable candidate for gas injection processes. As the achievement of miscibility leads to higher recoveries (see the comprehensive review on miscibility by Dindoruk et al., 2021), miscible hydrocarbon injection will be preferred where it is feasible. Using hydrocarbon gases for miscible injection has additional benefits 1) High outlet pressures from the separators (using gas compressors) (Griffith et al., 1970) leading to lesser gas volume to transport, 2) Reduction of flaring: in many places where there is no or not enough infrastructure capacity to transport these gases, such gases are simply flared (Worldbank, 2022; USCUSA, 2022), 3) Being associated with the in-situ fluids, less likely to have issues such as corrosion, or other potential problems. Methane (dominant component in hydrocarbon gas streams) and other live HC blends is classified greenhouse gases (GHG) which are harmful to the environment and such actions do not comply with the tighter carbon/emission reduction policies that are being implemented, especially considering the fact that resulting in the emission of methane, which has 28 to 36 times the global warming potential (GWP) compared to CO2 (EPA2, 2022; Turner et al., 2016; Turner et al., 2019). From the recovery point of view and as well as tuning the equation of state (EOS) for gas injection processes, minimum miscibility pressure (MMP) is a pivotal parameter estimating the recovery factor during gas injection processes. In this study we developed a methodology using Light Gradient Boost (Light GBM) (Ke et al., 2017) based model to estimate: 1) Minimum Miscibility Pressure (MMP) for the injected stream of hydrocarbon (HC) gas, and 2) minimum heavy HC gas (C2H6, C3H8 or their equivalence like CO2 and H2S) that is required to be added for a target MMP into the light HC gas stream. This is also known as minimum miscible enrichment, MME to achieve a prescribed MMP if the pressure requirements for a given lean gas are too high to implement in the field, usually due to compression cost and/or reservoir/well integrity reasons. We validate our MMP model using Leave-One-Out-Cross-Validation (LOOCV) (Shao, 1993) method which is the least unbiased cross -validation method and therefore proves the robustness of the proposed model. We compared the accuracy of the proposed model (AARE = 3.83%) with the leading correlations in the industry, and the proposed model significantly outperforms all the relevant models in the literature as shown in this study (Maklavani et al., 2010; Eakin and Mitch, 1988; Glaso, 1985; Kuo, 1985). The proposed model fully considers the underlying physics of miscibility and takes into consideration all the key and easy to obtain input parameters that influence the miscibility pressure of the injected gas.

A machine learning-based approach to ERα bioactivity and drug ADMET prediction.

By predicting ERα bioactivity and mining the potential relationship between Absorption, Distribution, Metabolism, Excretion, Toxicity (ADMET) attributes in drug research and development, the development efficiency of specific drugs for breast cancer will be effectively improved and the misjudgment rate of R&D personnel will be reduced. The quantitative prediction model of ERα bioactivity and classification prediction model of Absorption, Distribution, Metabolism, Excretion, Toxicity properties were constructed. The prediction results of ERα bioactivity were compared by XGBoot, Light GBM, Random Forest and MLP neural network. Two models with high prediction accuracy were selected and fused to obtain ERα bioactivity prediction model from Mean absolute error (MAE), mean squared error (MSE) and R2. The data were further subjected to model-based feature selection and FDR/FPR-based feature selection, respectively, and the results were placed in a voting machine to obtain Absorption, Distribution, Metabolism, Excretion, Toxicity classification prediction model. In this study, 430 molecular descriptors were removed, and finally 20 molecular descriptors with the most significant effect on biological activity obtained by the dual feature screening combined optimization method were used to establish a compound molecular descriptor prediction model for ERα biological activity, and further classification and prediction of the Absorption, Distribution, Metabolism, Excretion, Toxicity properties of the drugs were made. Eighty variables were selected by the model ExtraTreesClassifier Classifie, and 40 variables were selected by the model GradientBoostingClassifier to complete the model-based feature selection. At the same time, the feature selection method based on FDR/FPR is also selected, and the three classification models obtained by the two methods are placed into the voting machine to obtain the final model. The experimental results showed that the model's evaluation indexes and roc diagram were excellent and could accurately predict ERα bioactivity and Absorption, Distribution, Metabolism, Excretion, Toxicity properties. The model constructed in this study has high accuracy, fast convergence and robustness, has a very high accuracy for Absorption, Distribution, Metabolism, Excretion, Toxicity and ERα classification prediction, has bright prospects in the biopharmaceutical field, and is an important method for energy conservation and yield increase in the future.

Stacking Regression-Based Model for Predicting Patient's Length of Stay in a Semi Urban Hospital

In this research, the authors found that statistical analysis is very important preliminary phase in Machine Learning, especially for regression problems. Indeed, when the authors developed the first single models using the same algorithms and the same dataset, they obtained poor performances. After verifying the assumptions of the multiple linear regression, they adjusted the used data and produced efficient models. Moreover, as the objective was to apply the stacking model to predict Patient's Length of Stay in a semi urban hospital, the results showed that the stacking regressor performed better than the seven different models implemented (Random Forest, Extra Trees, Decision Tree, XGBoost, Multilayer perceptron, Light GBM, Support Vector Regressor (SVR)) taken individually. The authors combined Random Forest Regressor, Extra Trees Regressor, Decision Tree Regressor, XGBoost, Light GBM, and SVR to build the stacking model. Using secondary data from four services (Pediatrics, Hospitalization, Gynecology, and Neonatology) of a semi-urban hospital, located in a region of ongoing war in eastern Democratic Republic of Congo (DRC), the study examined the minimum length of stay of a patient in hospital when admitted in one of the four above services. Performances were evaluated using MAE, RMSE, MSE, R-squared and Accuracy. The stacking regression model shifted from 85% of accuracy before statistical analysis phase to 91% after applying statistics and from 0.75 to 0.91 as R-squared

Recognizing Safe Drinking Water and Predicting Water Quality Index using Machine Learning Framework

Water quality monitoring, analysis, and prediction have emerged as important challenges in several uses of water in our life. Recent water quality problems have raised the need for artificial intelligence (AI) models for analyzing water quality, classifying water samples, and predicting water quality index (WQI). In this paper, a machine-learning framework has been proposed for classify drinking water samples (safe/unsafe) and predicting water quality index. The classification tier of the proposed framework consists of nine machine-learning models, which have been applied, tested, validated, and compared for classifying drinking water samples into two classes (safe/unsafe) based on a benchmark dataset. The regression tier consists of six regression models that have been applied to the same dataset for predicting WQI. The experimental results clarified good classification results for the nine models with average accuracy, of 94.7%. However, the obtained results showed the superiority of Random Forest (RF), and Light Gradient Boosting Machine (Light GBM) models in recognizing safe drinking water samples regarding training and testing accuracy compared to the other models in the proposed framework. Moreover, the regression analysis results proved the superiority of LGBM regression, and Extra Trees Regression models in predicting WQI according to training, testing accuracy, 0.99%, and 0.95%, respectively. Moreover, the mean absolute error (MAE) results proved that the same models achieved less error rate, 10% than other applied regression models. These findings have significant implications for the understanding of how novel deep learning models can be developed for predicting water quality, which is suitable for other environmental and industrial purposes.

Shapley Values for Explaining the Black Box Nature of Machine Learning Model Clustering

Machine learning (ML) models are becoming increasingly complex. In fact, a sophisticated model (XGBoost boosting or deep learning) generally leads to more accurate predictions than a simple model (linear regression or decision tree). There is therefore a trade-off between the performance of a model and its interpretability: what a model gains in performance, it loses in interpretability (and vice versa), where interpretability is the ability for a human to understand the reasons for a model's decision. However, explaining the predictions made by machine learning models aims at computing and interpreting the importance of features. To achieve this, game theory has recently gained attention for better understanding the similarity between group members. In this paper, we use SHAP (SHapley Additive exPlanations), which is a method based on cooperative game theory, to analyze and evaluate the properties of each group. More importantly, we rely k-means PCA and Light gbm classifier to improve the data preparation before grouping the features into multiple clusters. The simulation results prove the importance of shapley value in creating an accurate and meaningful representation of each cluster.

428. Deep-Learning Based Predictive Model for Patients with Positive MRSA Cultures Using Time-Series Electronic Health Records

Abstract Background Methicillin-resistant Staphylococcus aureus (MRSA) is one of the common pathogens leading to significant morbidity and mortality in the hospital. This pathogen requires specific empirical antibiotics. Hence, identifying the personalized risks of this pathogen likely optimizes the usage of those antibiotics. Deep-learning based (DL) models are shown to be useful for modeling time-series electronic health record (EHR) data, and thus offering potential for predicting individualized risk for MRSA infection. Methods Longitudinal data were retrospectively retrieved from EHR in Memorial Hermann System, Houston, Tx. Patient encounters, demographics, diagnostic & procedure codes, antibiotics use, and microbiology data were extracted between 1/2018 and 4/2021. Inpatient and outpatient data were included. We randomly identified roughly equal numbers of patients with positive culture (Cx) for MRSA, MSSA, other pathogens, and negative Cx. We set a 2-week window for the prediction, and any first culture within the window was used as an index Cx. Some patients had multiple predictions over time and were included into both MRSA vs. non-MRSA groups. Our team developed a DL model platform (Pyorch_EHR) for clinical outcomes predictions using structured EHR data. In this project, the outcome is MRSA positivity taken within 2 weeks from the index Cx. Datasets are split into 50, 30, and 20% for training, validation, and test, respectively. Other machine learning models; logistic regression (LR) and light GBM (LGBM) were used for comparison. Pytorch ver. 1.7.1 and Sklearn ver. 0.24.2 are used. Results A total of 8164 patients and 22,563 patients were identified as MRSA and non-MRSA groups, respectively. Table 1 summarizes the key features of the patient’s characteristics. After model training using train and validation datasets, our model achieved an AUC of 91.8 in test dataset, whereas AUC of 86.0 and 88.2 in LR and LGBM, respectively. Figure 1 shows the cumulative incidence of 2 week MRSA positivity. Our model precisely stratified the risks. Figure 1:Cumulative Incidence of Positive MRSA Cultures Based on Risk Stratificationa Conclusion Our DL based model exhibited excellent performance in the prediction of 2 week MRSA positivity. Our model precisely categorized patients into low, medium, and high-risk. This work should be validated with other data sources and high-risk subgroups. Disclosures All Authors: No reported disclosures.

Predictive modeling the probability of suffering from metabolic syndrome using machine learning: A population-based study

BackgroundThere is an increasing trend of Metabolic syndrome (MetS) prevalence, which has been considered as an important contributor for cardiovascular disease (CVD), cancers and diabetes. However, there is often a long asymptomatic phase of MetS, resulting in not diagnosed and intervened so timely as needed. It would be very helpful to explore tools to predict the probability of suffering from MetS in daily life or routinely clinical practice. ObjectiveTo develop models that predict individuals’ probability of suffering from MetS timely with high efficacy in general population. MethodsThe present study enrolled 8964 individuals aged 40–75 years without severe diseases, which was a part of the REACTION study from October 2011 to February 2012. We developed three prediction models for different scenarios in hospital (Model 1, 2) or at home (Model 3) based on LightGBM (LGBM) technique and corresponding logistic regression (LR) models were also constructed for comparison. Model 1 included variables of laboratory tests, lifestyles and anthropometric measurements while model 2 was built with components of MetS excluded based on model 1, and model 3 was constructed with blood biochemical indexes removed based on model 2. Additionally, we also investigated the strength of association between the predictive factors and MetS, as well as that between the predictors and each component of MetS. ResultsIn this study, 2714 (30.3%) participants suffer from MetS accordingly. The performances of the LGBM models in predicting the probability of suffering from MetS produced good results and were presented as follows: model 1 had an area under the curve (AUC) value of 0.993 while model 2 indicated an AUC value of 0.885. Model 3 had an AUC value of 0.859, which is close to that of model 2. The AUC values of LR model 1 and 2 for the scenario in hospital and model 3 at home were 0.938, 0.839 and 0.820 respectively, which seemed lower than that of their corresponding machine learning models, respectively. In both LGBM and logistic models, gender, height and resting pulse rate (RPR) were predictors for MetS. Women had higher risk of MetS than men (OR 8.84, CI: 6.70–11.66), and each 1-cm increase in height indicated 3.8% higher risk of suffering from MetS in people over 58 years, whereas each 1- Beat Per Minute (bpm) increase in RPR showed 1.0% higher risk in individuals younger than 62 years. ConclusionThe present study showed that the prediction models developed by machine learning demonstrated effective in evaluating the probability of suffering from MetS, and presented prominent predicting efficacies and accuracies. Additionally, we found that women showed a higher risk of MetS than men, and height in individuals over 58 years was important factor in predicting the probability of suffering from MetS while RPR was of vital importance in people aged 40–62 years.

Automatically identifying the motivations of Facebook mention to scholarly outputs based on Light GBM

Automatically identifying the motivations of Facebook mention to scholarly outputs based on Light GBM

Modified light GBM based classification of malicious users in cooperative cognitive radio networks

ABSTRACT This paper proposes the Modified Light GBM to classify the Malicious Users (MUs) and legitimate Secondary Users (SUs) in the cognitive-radio network. The proposed method is to avoid the consequences of malicious users in Cooperative Spectrum Sensing (CSS) without the detection of MUs. The method is tested against the occurrence of Always No Malicious User (ANMU), always yes malicious user (AYMU), the Random Malicious User (RMU), and Opposite Malicious User (OMU) transmitting spectrum sensing data to the Fusion Center (FC) by normal secondary users. The efficiency of the proposed method is expressed via simulations and compared with other existing methods.

Detection of DDoS attacks in D2D communications using machine learning approach

In device-to-device (D2D) communications, distributed Denial-of-Service (DDoS) attacks can be quite detrimental because it can result in network structure destruction. Towards this end, the research objective of this paper is to identify and prevent DDoS and Denial-of-Service (DoS) attacks (i.e., SYN, Slowloris) in a D2D communication environment. Specifically, by replicating a real-world scenario, we emulate SLowloris attacks in a D2D communication network and generate a D2D Network-specific Slowloris dataset. This dataset along with the CICDDoS2019 dataset was then used to train our proposed Machine learning (ML) model that aids in the detection and prevention of DDoS attacks (Slowloris and SYN) in the considered D2D framework. The whole process of how to construct an emulation network for D2D communication and test it against a variety of attacks and implementations is also demonstrated in the paper. To quantify the detection accuracy in the context of DDoS and DoS attacks, we use various ML algorithms such as Random Forest, Light GBM, XGBoost, and Ada Boost and study their performance with the aid of extensive emulation. The results collected revealed that both Slowloris and CICDDoS2019 datasets achieve greater accuracy with Random Forest. Consequently, the results compel us to develop a technique for combining the identification of DDoS and DoS attacks in binary classification Random Forests with the binary decision. The proposed technique has been evaluated and compared with other related approaches in the open literature demonstrating significant performance in terms of identification and prevention time, processing and memory resources required, and device battery consumption, without affecting the accuracy of the attack identification. Hence, we advocate that our proposed technique can be extremely beneficial in preventing DDoS and DoS attacks in a D2D communication environment, where its lifetime and capabilities are mainly associated with the resources of the D2D device (i.e., CPU, Memory, and battery life).

Soil Classification by Machine Learning Using a Tunnel Boring Machine’s Operating Parameters

This study predicted soil classification using data gathered during the operation of an earth-pressure-balance-type tunnel boring machine (TBM). The prediction methodology used machine learning to find relationships between the TBM’s operating parameters which are monitored continuously during excavation, and the engineering characteristics of the ground which are only available from prior geotechnical investigation. Classification criteria were set using the No. 200 sieve pass rate and N-value and employed classification algorithms that used data for six operating parameters (penetration rate, thrust force, cutterhead torque, screw torque, screw revolution speed, and earth pressure). The results of the ensemble model (i.e., AdaBoost, gradient boosting, XG boosting, and Light GBM), decision tree, and SVM model were examined. As a result, the decision tree and AdaBoost models showed accuracy values of 0.759 to 0.879 in the first and second classification steps, but with poor precision and recall values of around 0.6. In contrast, the gradient boosting, XG boosting, Light GBM, and support vector models all showed excellent performance, with accuracy values over 0.90, and strong precision and recall values. Comparing the performance and the speed of learning using the same PC found Light GBM which showed both excellent learning performance and speed to be a suitable model for predicting soil classification using TBM operating data. The classification model developed here is expected to help guide excavation in sections of ground that lack prior geotechnical information.

Efficacy of a machine learning-based approach in predicting neurological prognosis of cervical spinal cord injury patients following urgent surgery within 24 h after injury

We aimed to develop a machine learning (ML) model for predicting the neurological outcomes of cervical spinal cord injury (CSCI). We retrospectively analyzed 135 patients with CSCI who underwent surgery within 24 h after injury. Patients were assessed with the American Spinal Injury Association Impairment Scale (AIS; grades A to E) 6 months after injury. A total of 34 features extracted from demographic variables, surgical factors, laboratory variables, neurological status, and radiological findings were analyzed. The ML model was created using Light GBM, XGBoost, and CatBoost. We evaluated Shapley Additive Explanations (SHAP) values to determine the variables that contributed most to the prediction models. We constructed multiclass prediction models for the five AIS grades and binary classification models to predict more than one-grade improvement in AIS 6 months after injury. Of the ML models used, CatBoost showed the highest accuracy (0.800) for the prediction of AIS grade and the highest AUC (0.90) for predicting improvement in AIS. AIS grade at admission, intramedullary hemorrhage, longitudinal extent of intramedullary T2 hyperintensity, and HbA1c were identified as important features for these prediction models. The ML models successfully predicted neurological outcomes 6 months after injury following urgent surgery in patients with CSCI.

A machine learning method for predicting the probability of MODS using only non-invasive parameters

Timely and accurate prediction of multiple organ dysfunction syndrome (MODS) is essential for the rescue and treatment of trauma patients However, existing methods are invasive, easily affected by artifacts and can be difficult to perform in a pre-hospital setting. We aim to develop prediction models for patients with MODS using only non-invasive parameters. In this study, records from 2319 patients were extracted from the Multiparameter Intelligent Monitoring in Intensive Care Ⅲ database (MIMIC Ⅲ), based on the sequential organ failure assessment (SOFA) score. Seven commonly used machine learning (ML) methods were selected and applied to develop a real-time prediction method for MODS based on full parameters (laboratory parameter. drug and non-invasive parameters, 57 parameters in total) and non-invasive parameters only (17 parameters) and compared with four traditional scoring systems. The prediction results using LightGBM (LGBM) and Adaboost based on the full parameter modeling were 0.959 for area under receiver operating characteristic curve (AUC), outperforming four traditional scoring systems. The removal of 40 parameters and retaining of 17 non-invasive parameters decreased the AUC value of LGBM by 0.015, which still outperformed all traditional scoring systems. A real-time and accurate MODS prediction method was developed in this paper based on non-invasive parameters by comparing the performance of four ML methods, which proved to be superior to the traditional scoring systems. This method can help medical staff to diagnose MODS as soon as possible and can improve the survival rate of patients in a pre-hospital setting.

Machine-learning models for prediction of sepsis patients mortality

ObjectivesSepsis is an infection-caused syndrome, that leads to life-threatening organ damage. We aim to develop machine learning models with large-scale data to predict sepsis patients’ mortality. Designwe extracted sepsis patients from two databases, Medical Information Mart for Intensive Care IV (MIMIC-IV) as a train set and Philips eICU Collaborative Research Database as a test set. SettingICUs in multicenter hospitals in the USA during 2012–2019. Patients or participantsA total of 21,680 sepsis-3 patients are included in the study, in which, 3771 patients were dead and 17,909 survived during hospitalization, respectively. InterventionsNo interventions. Main variables of interestBasic information, examination items during hospitalization and some medication and treatment information are incorporated into analyzed. Seven different models were built with a Support vector machine, Decision Tree Classifier, Random Forest, Gradients Boosting, Multiple Layer Perception, Xgboost, light Gradients Boosting to predict dead or live during hospitalization. ResultsAlgorithms with an AUC value in the test set of the top three: light GBM, GBM, Xgboost. Considering the performance of the training set and the test set, the light GBM model performs best, and then the parameters of the model were adjusted, after that the AUC value was 0.99 in the train set, 0.96 in the test set, respectively. ConclusionsModels built with light GBM algorithm from real-world sepsis patients from electronic health records accurately predict whether sepsis patients are dead and can be incorporated into clinical decision tools to enhance the prognosis of the patient and prevent adverse outcomes.

Investigating the Bond Strength of FRP Laminates with Concrete Using LIGHT GBM and SHAPASH Analysis.

The corrosion of steel reinforcement necessitates regular maintenance and repair of a variety of reinforced concrete structures. Retrofitting of beams, joints, columns, and slabs frequently involves the use of fiber-reinforced polymer (FRP) laminates. In order to develop simple prediction models for calculating the interfacial bond strength (IBS) of FRP laminates on a concrete prism containing grooves, this research evaluated the nonlinear capabilities of three ensemble methods-namely, random forest (RF) regression, extreme gradient boosting (XGBoost), and Light Gradient Boosting Machine (LIGHT GBM) models-based on machine learning (ML). In the present study, the IBS was the desired variable, while the model comprised five input parameters: elastic modulus x thickness of FRP (EfTf), width of FRP plate (bf), concrete compressive strength (fc'), width of groove (bg), and depth of groove (hg). The optimal parameters for each ensemble model were selected based on trial-and-error methods. The aforementioned models were trained on 70% of the entire dataset, while the remaining data (i.e., 30%) were used for the validation of the developed models. The evaluation was conducted on the basis of reliable accuracy indices. The minimum value of correlation of determination (R2 = 0.82) was observed for the testing data of the RF regression model. In contrast, the highest (R2 = 0.942) was obtained for LIGHT GBM for the training data. Overall, the three models showed robust performance in terms of correlation and error evaluation; however, the trend of accuracy was obtained as follows: LIGHT GBM > XGBoost > RF regression. Owing to the superior performance of LIGHT GBM, it may be considered a reliable ML prediction technique for computing the bond strength of FRP laminates and concrete prisms. The performance of the models was further supplemented by comparing the slopes of regression lines between the observed and predicted values, along with error analysis (i.e., mean absolute error (MAE), and root-mean-square error (RMSE)), predicted-to-experimental ratio, and Taylor diagrams. Moreover, the SHAPASH analysis revealed that the elastic modulus x thickness of FRP and width of FRP plate are the factors most responsible for IBS in FRP.

A novel hybrid method for oil price forecasting with ensemble thought

As a globally hard currency, crude oil has influenced all aspects of the world, especially because volatility in its price seriously affect economies, politics and wars. The prediction of oil prices is challenging due to remarkable price volatility, uncertainty, elusiveness, and complexity. There have been many papers adopting traditional machine learning (ML), economic approaches or combinations of the two; however, these papers have rarely focused on the performance of individual models and ignored less popular but complex models. Considering the above difficulties and situations, in this paper, we design an advanced approach for valuable and robust forecasting of crude oil prices to investigate the differences among the three other popular methods. For this purpose, the Random Forest (RF), XGBoost (XGB), and LightGBM (LGBM) are employed in the proposed hybrid method. The merits of this hybridization lie in the fact that the ensemble model is capable of handling volatile features such as nonlinearity, noncyclicity, and market interrelationships existing in oil price time series; incorporating the strengths of the three single models; not easily overfitting; and finally achieving a better performance. The results reveal that the hybrid method achieves the highest Directional Accuracy (DA) and the lowest errors (0.9612, 13.7417, and 0.0368 of DA, MAE, and MAPE, respectively) compared with the results of three other methods-the RF (0.9569, 14.2699, and 0.0385, respectively), XGB (0.9526, 14.5111, and 0.0387, respectively), and the LGBM (0.9526, 14.4022, and 0.0384, respectively)-in experiments on real datasets.

Improved Mayfly Optimization and LightGBM Classifier for Smart City Traffic Prediction

Objectives: This research work focuses on predicting traffic for the Smart City. Methods: Current research methods for traffic prediction are based on machine learning (ML) model. This article presents two contributions related to it. First, it provides feature engineering that includes feature extraction and a nature inspired optimization algorithm for selecting the best features. The mayfly optimization algorithm is improved by using the mode-based ranking method to select the best feature. Second, it uses the light-weight boosting method to train the datasets for better accuracy.Findings: The proposed Improved MayFly Optimization with LightGBM (IMFO-LGBM) is experimented with popular smart city datasets which is available in Kaggle website. IMFOLGBM shows an improvement in the prediction accuracy when compared with the baseline methods. It shows 2% of increase in the overall accuracy. Novelty:Nature inspired Mayfly optimization is improved and used to find the best feature for prediction. The selected features are then trained with the light weight boosting algorithm (i.e., Light Gradient Boosting Model). The hybrid of improved mayfly optimization and light GBM outperformed well. Keywords: IoT; Smartcity; Mayfly optimization; Machine learning and LightGBM

An interpretable machine learning prognostic system for risk stratification in oropharyngeal cancer

BackgroundThe optimal management of oropharyngeal squamous cell carcinoma (OPSCC) includes both surgical and non-surgical, that is, (chemo)radiotherapy treatment options and their combinations. These approaches carry a risk of specific treatment-related side effects. HPV-positive OPSCC has been reported to be more sensitive to (chemo)radiotherapy-based treatment modalities. Objectives: This study aims to demonstrate how machine learning can aid in classifying OPSCC patients into risk groups (low-chance or high-chance) for overall survival. We examined the input variables using permutation feature importance. Furthermore, we provided explanations and interpretations using the Local Interpretable Model Agnostic Explanations (LIME) and SHapley Additive Explanation (SHAP) frameworks. Methods: The machine learning model for 3164 OPSCC patients was built using data obtained from the Surveillance, Epidemiology, and End Results (SEER) program database. A total of five variants of tree-based machine learning algorithms (voting ensemble, light GBM, XGBoost, Random Forest, and Extreme Random Trees) were used to divide the patients into risk groups. The developed model with the best predictive performance was temporally validated with a different cohort. Results: The voting ensemble machine learning algorithm showed an accuracy of 88.3%, Mathews’ correlation coefficient of 0.72, and weighted area under curve of 0.93, when temporally validated. Human papillomavirus (HPV) status, age of the patients, T stage, marital status, N stage, and the treatment modality (surgery with postoperative radiotherapy) were found to have the most significant effects on the ability of the machine learning model to predict overall survival. Similarly, for the individual patients with SHAP framework, HPV status, gender, and treatment modality (surgery with postoperative radiotherapy) were the input features that improved the model’s prediction. Conclusion: The proposed stratification of OPSCC patients into risk groups by machine learning techniques can provide accurate predictions and thus aid clinicians in administering early and personalized interventions. Clinicians could utilize the predicted risk with the explanations offered by the SHAP and LIME frameworks to understand previously undetected relationships between prognostic variables to make informed clinical decisions and effective interventions.

The level of delay caused by crashes (LDC) in metropolitan and non-metropolitan areas: a comparative analysis of improved Random Forests and LightGBM

Traffic crashes cause serious traffic delay and have some unobserved heterogeneity in different areas. Using Texas accident data in 2020, this article aims to predict the level of delay caused by crashes (LDC) accurately and efficiently and discuss the difference between metropolitan and non-metropolitan areas. A framework based on Random Forests (RF) and LightGBM (LGBM) is developed to measure the association between LDC and its possible risk factors. At first, the most relevant variables in different areas were recognised through recursive feature elimination based on logistic regression. Then, LDC were forecasted by classifiers after grid search hyper parameters. To resolve data imbalance, two threshold moving methods of maximisation G-mean and F1-score were used. Finally, SHapley Additive explanation was employed to interpret the best model. The results indicate that the improved RF performs better in metropolitan areas and the improved LGBM performs better in non-metropolitan areas. In addition, Highway, spring and sunrise are the main risk factors of higher LDC in the two areas. And excessive wind speed and temperature in metropolitan areas can lead to higher LDC while in non-metropolitan areas it is pressure and apparent temperature.

Early Prediction of Diabetes Using an Ensemble of Machine Learning Models.

Diabetes is one of the most rapidly spreading diseases in the world, resulting in an array of significant complications, including cardiovascular disease, kidney failure, diabetic retinopathy, and neuropathy, among others, which contribute to an increase in morbidity and mortality rate. If diabetes is diagnosed at an early stage, its severity and underlying risk factors can be significantly reduced. However, there is a shortage of labeled data and the occurrence of outliers or data missingness in clinical datasets that are reliable and effective for diabetes prediction, making it a challenging endeavor. Therefore, we introduce a newly labeled diabetes dataset from a South Asian nation (Bangladesh). In addition, we suggest an automated classification pipeline that includes a weighted ensemble of machine learning (ML) classifiers: Naive Bayes (NB), Random Forest (RF), Decision Tree (DT), XGBoost (XGB), and LightGBM (LGB). Grid search hyperparameter optimization is employed to tune the critical hyperparameters of these ML models. Furthermore, missing value imputation, feature selection, and K-fold cross-validation are included in the framework design. A statistical analysis of variance (ANOVA) test reveals that the performance of diabetes prediction significantly improves when the proposed weighted ensemble (DT + RF + XGB + LGB) is executed with the introduced preprocessing, with the highest accuracy of and an area under the ROC curve (AUC) of . In conjunction with the suggested ensemble model, our statistical imputation and RF-based feature selection techniques produced the best results for early diabetes prediction. Moreover, the presented new dataset will contribute to developing and implementing robust ML models for diabetes prediction utilizing population-level data.