Accuracy Of Machine Learning Models Research Articles

An Efficient Deep Learning Approach for DNA-Binding Proteins Classification from Primary Sequences

As the number of identified proteins has expanded, the accurate identification of proteins has become a significant challenge in the field of biology. Various computational methods, such as Support Vector Machine (SVM), K-nearest neighbors (KNN), and convolutional neural network (CNN), have been proposed to recognize deoxyribonucleic acid (DNA)-binding proteins solely based on amino acid sequences. However, these methods do not consider the contextual information within amino acid sequences, limiting their ability to adequately capture sequence features. In this study, we propose a novel approach to identify DNA-binding proteins by integrating a CNN with bidirectional long-short-term memory (LSTM) and gated recurrent unit (GRU) as (CNN-BiLG). The CNN-BiLG model can explore the potential contextual relationships of amino acid sequences and obtain more features than traditional models. Our experimental results demonstrate a validation set prediction accuracy of 94% for the proposed CNN-BiLG, surpassing the accuracy of machine learning models and deep learning models. Furthermore, our model is both effective and efficient, exhibiting commendable classification accuracy based on comparative analysis.

The improvement of ground truth annotation in public datasets for human detection

The quality of annotations in the datasets is crucial for supervised machine learning as it significantly affects the performance of models. While many public datasets are widely used, they often suffer from annotations errors, including missing annotations, incorrect bounding box sizes, and positions. It results in low accuracy of machine learning models. However, most researchers have traditionally focused on improving model performance by enhancing algorithms, while overlooking concerns regarding data quality. This so-called model-centric AI approach has been predominant. In contrast, a data-centric AI approach, advocated by Andrew Ng at the DATA and AI Summit 2022, emphasizes enhancing data quality while keeping the model fixed, which proves to be more efficient in improving performance. Building upon this data-centric approach, we propose a method to enhance the quality of public datasets such as MS-COCO and Open Image Dataset. Our approach involves automatically retrieving missing annotations and correcting the size and position of existing bounding boxes in these datasets. Specifically, our study deals with human object detection, which is one of the prominent applications of artificial intelligence. Experimental results demonstrate improved performance with models such as Faster-RCNN, EfficientDet, and RetinaNet. We can achieve up to 32% compared to original datasets in the term of mAP after applying both proposed methods to dataset which is transformed the grouped of instances to individual instance. In summary, our methods significantly enhance the model’s performance by improving the quality of annotations at a lower cost with less time than manual improvement employed in other studies.

Improving permeability prediction via Machine Learning in a heterogeneous carbonate reservoir: application to Middle Miocene Nullipore, Ras Fanar field, Gulf of Suez, Egypt

Predicting and interpolating the permeability between wells to obtain the 3D distribution is a challenging mission in reservoir simulation. The high degree of heterogeneity and diagenesis in the Nullipore carbonate reservoir provide a significant obstacle to accurate prediction. Moreover, intricate relationships between core and well logging data exist in the reservoir. This study presents a novel approach based on Machine Learning (ML) to overcome such difficulties and build a robust permeability predictive model. The main objective of this study is to develop an ML-based permeability prediction approach to predict permeability logs and populate the predicted logs to obtain the 3D permeability distribution of the reservoir. The methodology involves grouping the reservoir cored intervals into flow units (FUs), each of which has distinct petrophysical characteristics. The probability density function is used to investigate the relationships between the well logs and FUs to select high-weighted input features for reliable model prediction. Five ML algorithms, including Linear Regression (LR), Polynomial Regression (PR), Support Vector Regression (SVR), Decision Trees (DeT), and Random Forests (RF), have been implemented to integrate the core permeability with the influential well logs to predict permeability. The dataset is randomly split into training and testing sets to evaluate the performance of the developed models. The models’ hyperparameters were tuned to improve the model’s prediction performance. To predict permeability logs, two key wells containing the whole reservoir FUs are used to train the most accurate ML model, and other wells to test the performance. Results indicate that the RF model outperforms all other ML models and offers the most accurate results, where the adjusted coefficient of determination (R2adj) between the predicted permeability and core permeability is 0.87 for the training set and 0.82 for the testing set, mean absolute error and mean squared error (MSE) are 0.32 and 0.19, respectively, for both sets. It was observed that the RF model exhibits high prediction performance when it is trained on wells containing the whole reservoir FUs. This approach aids in detecting patterns between the well logs and permeability along the profile of wells and capturing the wide permeability distribution of the reservoir. Ultimately, the predicted permeability logs were populated via the Gaussian Random Function Simulation geostatistical method to build a 3D permeability distribution for the reservoir. The study outcomes will aid users of ML to make informed choices on the appropriate ML algorithms to use in carbonate reservoir characterization for more accurate permeability predictions and better decision-making with limited available data.

Interpretable Machine Learning for Finding Intermediate-mass Black Holes

Definitive evidence that globular clusters (GCs) host intermediate-mass black holes (IMBHs) is elusive. Machine-learning (ML) models trained on GC simulations can in principle predict IMBH host candidates based on observable features. This approach has two limitations: first, an accurate ML model is expected to be a black box due to complexity; second, despite our efforts to simulate GCs realistically, the simulation physics or initial conditions may fail to reflect reality fully. Therefore our training data may be biased, leading to a failure in generalization to observational data. Both the first issue—explainability/interpretability—and the second—out of distribution generalization and fairness—are active areas of research in ML. Here we employ techniques from these fields to address them: we use the anchors method to explain an Extreme Gradient Boosting (XGBoost) classifier; we also independently train a natively interpretable model using Certifiably Optimal RulE ListS (CORELS). The resulting model has a clear physical meaning, but loses some performance with respect to XGBoost. We evaluate potential candidates in real data based not only on classifier predictions but also on their similarity to the training data, measured by the likelihood of a kernel density estimation model. This measures the realism of our simulated data and mitigates the risk that our models may produce biased predictions by working in extrapolation. We apply our classifiers to real GCs, obtaining a predicted classification, a measure of the confidence of the prediction, an out-of-distribution flag, a local rule explaining the prediction of XGBoost, and a global rule from CORELS.

Detection of Autism Spectrum Disorder using Machine Learning: A Review

Abstract: Social interaction and communication impairments are caused by an illness known as Autism Spectrum Disorder (ASD), which has neurological and genetic components. Statistics from the World Health Organization (WHO) show that the number of people with ASD diagnoses is progressively rising. The majority of recent research focuses on data gathering, brain image analysis, and clinical diagnosis; it does not address the diagnosis of ASD using machine learning. Currently, the only techniques available for diagnosing ASD are clinical standardized testing. This results in longer diagnostic times as well as a sharp rise in medical expenses. Machine learning approaches are being used to supplement traditional methods in order to enhance diagnosis precision and time required. The dataset is being subjected to several machine learning approaches, with the aim of constructing predictive models that are contingent on the results. The best accurate machine learning model to identify the disease in its early stages is then found by analyzing each technique according to its assessment metrics.

A hybrid approach of anaerobic digestion model no. 1 and machine learning to model and optimize continuous anaerobic digestion processes

Anaerobic digestion is a promising approach to dispose of biodegradable waste and wastewater, generating biogas as an alternative energy resource. This work proposed a so-called M-CADM1 for continuous anaerobic digestion simulation, which combined the machine learning and anaerobic digestion model No.1 (ADM1). The detailed reaction path and intermediate products in different stages of anaerobic digestion are specified in ADM1. The kinetic parameters were modified by machine learning. The characteristics (elemental composition) of feedstocks were used to predict kinetic parameters. A total of 75 biomass samples were used to establish for machine learning models. Five element contents (C, H, O, N, S), feedstock feed rate, and anaerobic digestion temperature were used as the input. The kinetic parameters were set as output. The sensitivities of 17 kinetic parameters were evaluated. 7 kinetic parameters with the highest sensitivities were selected as ADM1 model inputs by sensitivity analysis. The R2 and RMSE were used as the index to evaluated the accuracy of machine learning model. The best R2 and RMSE reached 0.84 and 0.196. The TIC was used as the index to evaluated the accuracy of M-CADM1. By comparing the simulated value with the experimental value, the accuracy of the overall M-CADM1 expressed by TIC of kitchen waste was 0.036. The organic acid content and pH in the reactor were considered as indicators to study the accuracy and stability of the M-CADM1. Trends in organic acids, free ammonia or hydrogen inhibition, and pH were consistent with experimental continuous anaerobic digestion results.

Predicting acute myocardial infarction from haematological markers utilizing machine learning and explainable artificial intelligence

Myocardial infarction (MI) is the leading cause of human death globally. It occurs when a blockage in an artery prevents blood and oxygen from reaching the heart muscle, causing tissues in the heart muscle to die. This leads to a necessity to develop a method to diagnose MI’s early, preventing further complications such as irregular heart rhythm, heart failure or even cardiac arrest. This research aims to develop a more accurate machine learning (ML) model to help predict acute myocardial infarction (AMI) with a greater degree of accuracy without invasive procedures using additional explainable artificial intelligence (XAI) techniques which will help medical practitioners to better diagnose AMI more precisely. According to the results, the random forest classifier model gave the highest accuracy of 86%. XAI techniques were used to visualize the data and results, and determined white blood cell (WBC) count to be the most crucial feature in classification, followed by neutrophil (NEU) count, neutrophil-lymphocyte (NEU/LY) ratio, platelet width of distribution (PDW) values and basophil (BA) counts. The developed model can help medical practitioners make a more accurate early diagnosis of AMI using readily available hematological parameters, enabling practitioners to provide superior care to a diverse range of individuals.

DrFuse: Learning Disentangled Representation for Clinical Multi-Modal Fusion with Missing Modality and Modal Inconsistency

The combination of electronic health records (EHR) and medical images is crucial for clinicians in making diagnoses and forecasting prognoses. Strategically fusing these two data modalities has great potential to improve the accuracy of machine learning models in clinical prediction tasks. However, the asynchronous and complementary nature of EHR and medical images presents unique challenges. Missing modalities due to clinical and administrative factors are inevitable in practice, and the significance of each data modality varies depending on the patient and the prediction target, resulting in inconsistent predictions and suboptimal model performance. To address these challenges, we propose DrFuse to achieve effective clinical multi-modal fusion. It tackles the missing modality issue by disentangling the features shared across modalities and those unique within each modality. Furthermore, we address the modal inconsistency issue via a disease-wise attention layer that produces the patient- and disease-wise weighting for each modality to make the final prediction. We validate the proposed method using real-world large-scale datasets, MIMIC-IV and MIMIC-CXR. Experimental results show that the proposed method significantly outperforms the state-of-the-art models.

Analyzing Dataset Annotation Quality Management in the Wild

Abstract Data quality is crucial for training accurate, unbiased, and trustworthy machine learning models as well as for their correct evaluation. Recent works, however, have shown that even popular datasets used to train and evaluate state-of-the-art models contain a non-negligible amount of erroneous annotations, biases, or artifacts. While practices and guidelines regarding dataset creation projects exist, to our knowledge, large-scale analysis has yet to be performed on how quality management is conducted when creating natural language datasets and whether these recommendations are followed. Therefore, we first survey and summarize recommended quality management practices for dataset creation as described in the literature and provide suggestions for applying them. Then, we compile a corpus of 591 scientific publications introducing text datasets and annotate it for quality-related aspects, such as annotator management, agreement, adjudication, or data validation. Using these annotations, we then analyze how quality management is conducted in practice. A majority of the annotated publications apply good or excellent quality management. However, we deem the effort of 30% of the works as only subpar. Our analysis also shows common errors, especially when using inter-annotator agreement and computing annotation error rates.

Qualitative data augmentation for performance prediction in VLSI circuits

Various studies have shown the advantages of using Machine Learning (ML) techniques for analog and digital IC design automation and optimization. Data scarcity is still an issue for electronic designs, while training highly accurate ML models. This work proposes generating and evaluating artificial data using generative adversarial networks (GANs) for circuit data to aid and improve the accuracy of ML models trained with a small training data set. The training data is obtained by various simulations in the Cadence Virtuoso, HSPICE, and Microcap design environment with TSMC 180 nm and 22 nm CMOS technology nodes. The artificial data is generated and tested for an appropriate set of analog circuits and digital cells. The experimental results show that the proposed artificial data generation significantly improves ML models and reduces the percentage error by more than 50% of the original percentage error, which were previously trained with insufficient data. Furthermore, this research aims to contribute to the extensive application of AI/ML in the field of VLSI design and technology by relieving the training data availability-related challenges.

PREDICTION OF CONCRETE MIXTURE DESIGN AND COMPRESSIVE STRENGTH THROUGH DATA ANALYSIS AND MACHINE LEARNING

Concrete is the most used building material in civil engineering. The mechanical properties of concrete depend on the percentage of materials used in the mix design. There are different types of mixture methods, and the purpose of this study is to investigate the mechanical properties of concrete using the mixture method through data analysis. In this case, more than 45 mixture designs are collected to find the estimated mixture design. The estimated mixture design was found by correlation matrix and the correlation between materials of concrete. Moreover, to find the reliability of the compressive strength of concrete through data mining, two models have been established. In this term, Linear Regression (LR), Ridge Regression (RR), Support Vector Machine Regression (SVR), and Polynomial Regression (PR) have been applied to predict compressive strength. In this study, the stress-strain curve of the compressive strength of concrete was also investigated. To find the accuracy of machine learning models, Correlation Coefficient (R2), Mean Absolute Errors (MAE), and Root Mean Squared Errors (RMSE) are established. However, the machine learning prediction model of RR and PR shows the best results of prediction with R2 0.93, MAE 3.7, and RMSE 5.3 for RR. The PR R2 was more than 0.91, moreover, the stress-strain of compressive strengths has been predicted with high accuracy through Logistic Algorithm Function. The experimental results were acceptable. In the compressive strength experimental results R2 was 0.91 MAE was 1.07, and RMSE was 2.71 from prediction mixture designs. Finally, the prediction and experimental results have indicated that the current study was reliable.

Empowering prediction of miRNA–mRNA interactions in species with limited training data through transfer learning

MicroRNAs (miRNAs) play a crucial role in mRNA regulation. Identifying functionally important mRNA targets of a specific miRNA is essential for uncovering its biological function and assisting miRNA–based drug development. Datasets of high-throughput direct bona fide miRNA–target interactions (MTIs) exist only for a few model organisms, prompting the need for computational prediction. However, the scarcity of data poses a challenge in training accurate machine learning models for MTI prediction. In this study, we explored the potential of transfer learning technique (with ANN and XGB) to address the limited data challenge by leveraging the similarities in interaction rules between species. Furthermore, we introduced a novel approach called TransferSHAP for estimating the feature importance of transfer learning in tabular dataset tasks. We demonstrated that transfer learning improves MTI prediction accuracy for species with limited datasets and identified the specific interaction features the models employed to transfer information across different species.

Assessing the performance of state-of-the-art machine learning algorithms for predicting electro-erosion wear in cryogenic treated electrodes of mold steels

In manufacturing, predicting and reducing electro-erosion wear during the electric discharge machining (EDM) process is critical to minimize delays, financial losses and product defects. Achieving this requires developing and evaluating accurate machine learning models. In our study, we focus on cryogenically treated mold steel electrodes to investigate the potential of different machine learning algorithms to predict EDM wear. We considered five machine learning algorithms—artificial neural networks, ensemble learning, boosting algorithms, tree-based algorithms, and k-nearest neighbors—to evaluate their ability to predict wear patterns accurately. Each algorithm was trained and tested using actual experimental data from EDM processes. Our results show that the machine learning models demonstrated exceptional accuracy, accurately predicting EDM wear in training and testing datasets with almost 99% accuracy. In addition, we identified the most influential characteristics that affect wear patterns, including operating current, cryogenic process parameters, and electrode composition. Based on these findings, manufacturers can gain valuable insight into the factors that cause EDM wear and optimize their EDM processes accordingly to improve productivity, reduce wear-related costs, and increase production quality across multiple manufacturing industries. Furthermore, this research provides insights into the possibilities of implementing these models in real manufacturing contexts and motivates future research on this topic. Ultimately, integrating advanced computing techniques and prudent decision-making strategies will shape the future of manufacturing operations management and promote sustainable and profitable business growth.

Exploring an accurate machine learning model to quickly estimate stability of diverse energetic materials

High energy and low sensitivity have been the focus of developing new energetic materials (EMs). However, there has been a lack of a quick and accurate method for evaluating the stability of diverse EMs. Here, we develop a machine learning prediction model with high accuracy for bond dissociation energy (BDE) of EMs. A reliable and representative BDE dataset of EMs is constructed by collecting 778 experimental energetic compounds and quantum mechanics calculation. To sufficiently characterize the BDE of EMs, a hybrid feature representation is proposed by coupling the local target bond into the global structure characteristics. To alleviate the limitation of the low dataset, pairwise difference regression is utilized as a data augmentation with the advantage of reducing systematic errors and improving diversity. Benefiting from these improvements, the XGBoost model achieves the best prediction accuracy with R2 of 0.98 and MAE of 8.8kJ mol-1, significantly outperforming competitive models.

Implementation of Polynomial Functions to Improve the Accuracy of Machine Learning Models in Predicting the Corrosion Inhibition Efficiency of Pyridine-Quinoline Compounds as Corrosion Inhibitors

Historically, the exploration of corrosion inhibitor technology has relied extensively on experimental methodologies, which are inherently associated with substantial costs, prolonged durations, and significant resource utilization. However, the emergence of ML approaches has recently garnered attention as a promising avenue for investigating potential materials with corrosion inhibition properties. This study endeavors to enhance the predictive capacity of ML models by leveraging polynomial functions. Specifically, the investigation focuses on assessing the effectiveness of pyridinequinoline compounds in mitigating corrosion. Diverse ML models were systematically evaluated, integrating polynomial functions to augment their predictive capabilities. The integration of polynomial functions notably amplifies the predictive accuracy across all tested models. Notably, the SVR model emerges as the most adept, exhibiting R² of 0.936 and RMSE of 0.093. The outcomes of this inquiry underscore a significant enhancement in predictive accuracy facilitated by the incorporation of polynomial functions within ML models. The proposed SVR model stands out as a robust tool for prognosticating the corrosion inhibition potential of pyridine-quinoline compounds. This pioneering approach contributes invaluable insights into advancing machine learning methodologies geared toward designing and engineering materials with promising corrosion inhibition properties.
 Keywords: machine learning, polynomial, corrosion inhibition, pyridine-quinoline

Identifying the most accurate machine learning classification technique to detect network threats

AbstractInsider threats have recently become one of the most urgent cybersecurity challenges facing numerous businesses, such as public infrastructure companies, major federal agencies, and state and local governments. Our purpose is to find the most accurate machine learning (ML) model to detect insider attacks. In the realm of machine learning, the most convenient classifier is usually selected after further evaluation trials of candidate models which can cause unseen data (test data set) to leak into models and create bias. Accordingly, overfitting occurs because of frequent training of models and tuning hyperparameters; the models perform well on the training set while failing to generalize effectively to unseen data. The validation data set and hyperparameter tuning are utilized in this study to prevent the issues mentioned above and to choose the best model from our candidate models. Furthermore, our approach guarantees that the selected model does not memorize data of the threats occurring in the local area network (LAN) through the usage of the NSL-KDD data set. The following results are gathered and analyzed: support vector machine (SVM), decision tree (DT), logistic regression (LR), adaptive boost (AdaBoost), gradient boosting (GB), random forests (RFs), and extremely randomized trees (ERTs). After analyzing the findings, we conclude that the AdaBoost model is the most accurate, with a DoS of 99%, a probe of 99%, access of 96%, and privilege of 97%, as well as an AUC of 0.992 for DoS, 0.986 for probe, 0.952 for access, and 0.954 for privilege.

Assessing the knot in a noose position by thyrohyoid and cervical spine fracture patterns in suicidal hangings using machine learning algorithms: A new insight into old dilemmas

Hanging is one of the most common suicide methods worldwide. Neck injuries that occur upon such neck compression – fractures of the thyrohyoid complex and cervical spine, occupy forensic pathologists for a long time. However, research failed to identify particular patterns of these injuries corresponding to the force distribution a ligature applies to the neck: the issue of reconstructing the knot in a noose position persists. So far, machine learning (ML) models were not utilized to classify knot positions and reconstruct this event. We conducted a single-institutional, retrospective study on 1235 autopsy cases of suicidal hanging, developed several ML models, and assessed their classification performance in a stepwise manner to discriminate between: 1. typical (‘posterior) and atypical (‘anterior’ and ‘lateral’) hangings, 2. anterior and lateral hangings, and 3. left and right lateral hangings. The variable coding was based on the presence/absence of fractures of greater hyoid bone horns (GHH), superior thyroid cartilage horns (STH), and cervical spine. Subject age was considered. The models’ parameters were optimized by the Genetic Algorithm. The accuracy of ML models in the first step was very modest (c. 60%) but increased subsequently: Multilayer Perceptron – Artificial Neural Network and k-Nearest Neighbors performed excellently discriminating between left and right lateral hangings (accuracy 91.8% and 90.6%, respectively). The latter is of great importance for clarifying probable hanging fracture biomechanics. Alongside the conventional inferential statistical analysis we performed, our results further indicate the association of the knot position with ipsilateral GHH and contralateral STH fractures in lateral hangings. Moreover, odds for unilateral GHH fracture, simultaneous GHH and STH fractures, and cervical spine fracture were significantly higher in atypical (‘anterior’ and ‘lateral’) hangings, compared to typical (‘posterior’) hangings.

Energy or Accuracy? Near-Optimal User Selection and Aggregator Placement for Federated Learning in MEC

To unveil the hidden value in the datasets of user equipments (UEs) while preserving user privacy, federated learning (FL) is emerging as a promising technique to train a machine learning model using the datasets of UEs locally without uploading the datasets to a central location. Customers require to train machine learning models based on different datasets of UEs, through issuing FL requests that are implemented by FL services in a mobile edge computing (MEC) network. A key challenge of enabling FL in MEC networks is how to minimize the energy consumption of implementing FL requests while guaranteeing the accuracy of machine learning models, given that the availabilities of UEs usually are uncertain. In this paper, we investigate the problem of energy minimization for FL in an MEC network with uncertain availabilities of UEs. We first consider the energy minimization problem for a single FL request in an MEC network. We then propose a novel optimization framework for the problem with a single FL request, which consists of (1) an online learning algorithm with a bounded regret for the UE selection, by considering various contexts (side information) that influence energy consumption; and (2) an approximation algorithm with an approximation ratio for the aggregator placement for a single FL request. We thirdly deal with the problem with multiple FL requests, for which we devise an online learning algorithm with a bounded regret. We finally evaluate the performance of the proposed algorithms by extensive experiments. Experimental results show that the proposed algorithms outperform their counterparts by reducing at least 13% of the total energy consumption while achieving the same accuracy.

Fair and green hyperparameter optimization via multi-objective and multiple information source Bayesian optimization

AbstractIt has been recently remarked that focusing only on accuracy in searching for optimal Machine Learning models amplifies biases contained in the data, leading to unfair predictions and decision supports. Recently, multi-objective hyperparameter optimization has been proposed to search for Machine Learning models which offer equally Pareto-efficient trade-offs between accuracy and fairness. Although these approaches proved to be more versatile than fairness-aware Machine Learning algorithms—which instead optimize accuracy constrained to some threshold on fairness—their carbon footprint could be dramatic, due to the large amount of energy required in the case of large datasets. We propose an approach named FanG-HPO: fair and green hyperparameter optimization (HPO), based on both multi-objective and multiple information source Bayesian optimization. FanG-HPO uses subsets of the large dataset to obtain cheap approximations (aka information sources) of both accuracy and fairness, and multi-objective Bayesian optimization to efficiently identify Pareto-efficient (accurate and fair) Machine Learning models. Experiments consider four benchmark (fairness) datasets and four Machine Learning algorithms, and provide an assessment of FanG-HPO against both fairness-aware Machine Learning approaches and two state-of-the-art Bayesian optimization tools addressing multi-objective and energy-aware optimization.

Predicting major adverse cardiovascular events after orthotopic liver transplantation using a supervised machine learning model: A cohort study.

Liver transplant (LT) patients have become older and sicker. The rate of post-LT major adverse cardiovascular events (MACE) has increased, and this in turn raises 30-d post-LT mortality. Noninvasive cardiac stress testing loses accuracy when applied to pre-LT cirrhotic patients. To assess the feasibility and accuracy of a machine learning model used to predict post-LT MACE in a regional cohort. This retrospective cohort study involved 575 LT patients from a Southern Brazilian academic center. We developed a predictive model for post-LT MACE (defined as a composite outcome of stroke, new-onset heart failure, severe arrhythmia, and myocardial infarction) using the extreme gradient boosting (XGBoost) machine learning model. We addressed missing data (below 20%) for relevant variables using the k-nearest neighbor imputation method, calculating the mean from the ten nearest neighbors for each case. The modeling dataset included 83 features, encompassing patient and laboratory data, cirrhosis complications, and pre-LT cardiac assessments. Model performance was assessed using the area under the receiver operating characteristic curve (AUROC). We also employed Shapley additive explanations (SHAP) to interpret feature impacts. The dataset was split into training (75%) and testing (25%) sets. Calibration was evaluated using the Brier score. We followed Transparent Reporting of a Multivariable Prediction Model for Individual Prognosis or Diagnosis guidelines for reporting. Scikit-learn and SHAP in Python 3 were used for all analyses. The supplementary material includes code for model development and a user-friendly online MACE prediction calculator. Of the 537 included patients, 23 (4.46%) developed in-hospital MACE, with a mean age at transplantation of 52.9 years. The majority, 66.1%, were male. The XGBoost model achieved an impressive AUROC of 0.89 during the training stage. This model exhibited accuracy, precision, recall, and F1-score values of 0.84, 0.85, 0.80, and 0.79, respectively. Calibration, as assessed by the Brier score, indicated excellent model calibration with a score of 0.07. Furthermore, SHAP values highlighted the significance of certain variables in predicting postoperative MACE, with negative noninvasive cardiac stress testing, use of nonselective beta-blockers, direct bilirubin levels, blood type O, and dynamic alterations on myocardial perfusion scintigraphy being the most influential factors at the cohort-wide level. These results highlight the predictive capability of our XGBoost model in assessing the risk of post-LT MACE, making it a valuable tool for clinical practice. Our study successfully assessed the feasibility and accuracy of the XGBoost machine learning model in predicting post-LT MACE, using both cardiovascular and hepatic variables. The model demonstrated impressive performance, aligning with literature findings, and exhibited excellent calibration. Notably, our cautious approach to prevent overfitting and data leakage suggests the stability of results when applied to prospective data, reinforcing the model's value as a reliable tool for predicting post-LT MACE in clinical practice.