Machine Learning Studies Research Articles

Advanced Machine Learning and Experimental Studies of Polypropylene Based Polyesters Tribological Composite Systems for Sustainable Recycling Automation and Digitalization

Digitalization and automation are emerging solutions to the complex problems of recycling. In this research work, the experimental and Python based Archard deep learning wear rate models are introduced regarding recycling automation and composite tribological systems optimization. The optimum polyester fibers (PESF) of length of 3-3.5 mm were used for fabrication of polypropylene (PP)-PESF composite systems. The deformation, high texture, asperities, and micro-cracks were observed during scanning electron microscope and machine-learning studies. The lowest experimental value of abrasive wear of 3.0 × 10-6 mm3/Nm was observed for PP. Comparatively, higher experimental values of abrasive wear of the PP-PESF composites are found in the range of 4.35 × 10-6 to 4.7 × 10-6 mm3/Nm due to presence micro-defects on the surface of composites. The experimental values of Coefficient of friction (COF) of PP and PP-PESF are found in the range of 0.70 - 0.8 and 1.1 - 1.3, respectively. The experimental values of abrasive wear and COF are found compatible with literature. Similarly, the simulated values of abrasive wear of PP and PP-PESF composites are predicted in the range of 4.8×10-7 to 3.75×10-7 mm3/Nm, respectively. The predicted values of PP and PP-PESF composite show better resistance towards abrasive wear. The proposed experimental and simulated (in terms of Python coding, machine learning, image processing, artificial intelligence, and deep learning studies) research work can be introduced industrially for automation as well as digitalization of grinding of PES waste, processing, tribological testing, and SEM characterization evaluations.

Is an Electronic Nose Able to Predict Clinical Response following Neoadjuvant Treatment of Rectal Cancer? A Prospective Pilot Study.

Introduction: A watch-and-wait strategy for patients with rectal cancer who achieve a clinical complete response after neoadjuvant (chemo) radiotherapy is a valuable alternative to rectal resection. In this pilot study, we explored the use of an electronic nose to predict response to neoadjuvant therapy by analyzing breath-derived volatile organic compounds. Materials and Methods: A pilot study was performed between 2020 and 2022 on patients diagnosed with intermediate- or high-risk rectal cancer who were scheduled for neoadjuvant therapy. Breath samples were collected before and after (chemo) radiotherapy. A machine-learning model was developed to predict clinical response using curatively treated rectal cancer patients as controls. Results: For developing the machine-learning model, a total of 99 patients were included: 45 patients with rectal cancer and 54 controls. In the training set, the model successfully discriminated between patients with and without rectal cancer, with a sensitivity and specificity of 0.80 and 0.65, respectively, and an accuracy of 0.72. In the test set, the model predicted partial or (near) complete response with a sensitivity and specificity of 0.64 and 0.47, respectively, and an accuracy of 0.58. The AUC of the ROC curve was 0.63. Conclusions: The prediction model developed in this pilot study lacks the ability to accurately differentiate between partial and (near) complete responders with an electronic nose. Machine-learning studies demand a substantial number of patients and operate in a rapidly evolving field. Therefore, the prevalence of disease and duration of a study are crucial considerations for future research.

Predictive modelling of Immunogenicity to Botulinumtoxin A Treatments for Glabellar Lines.

Botulinum toxin A (BoNT-A), derived from Clostridium botulinum, is widely used in medical and aesthetic treatments. Its clinical application extends from managing chronic conditions like cervical dystonia and migraine to reducing facial wrinkles. Despite its efficacy, a significant challenge associated with BoNT-A therapy is immunogenicity, where the immune system produces neutralising antibodies (NAbs) against BoNT-A, reducing its effectiveness over time. This issue is critical for patients requiring repeated treatments. The study aims to compare FDA-approved BoNT-A products, examining the factors influencing NAbs development using advanced machine learning techniques. This research analysed data from randomised controlled trials (RCTs) involving five main BoNT-A products. The study selected trials based on detailed immunogenic responses to these treatments, particularly for glabellar lines. Machine learning models, including logistic regression, random forest classifiers, and Bayesian logistic regression, were employed to assess how treatment specifics and BoNT-A product types affect the development of NAbs. Analysis of 14 studies with 8,190 participants revealed that dosage and treatment frequency are key factors influencing the risk of NAbs development. Among BoNT-A products, IncobotulinumtoxinA shows the lowest, and AbobotulinumtoxinA the highest likelihood of inducing NAbs. The study's machine learning and logistic regression findings indicated that treatment planning must consider these variables to minimise immunogenicity. The study underscores the importance of understanding BoNT-A immunogenicity in clinical practice. By identifying the main predictors of NAbs development and differentiating the immunogenic potential of BoNT-A products, the research provides insights for clinicians in optimising treatment strategies. It highlights the need for careful treatment customisation to reduce immunogenic risks, advocating for further research into the mechanisms of BoNT-A immunogenicity.

Machine Learning-Inspired Molecular Design, Divergent Syntheses, and X-Ray Analyses of Dithienobenzothiazole-Based Semiconductors Controlled by S⋅⋅⋅N and S⋅⋅⋅S Interactions.

Inspired by the previous machine-learning study that the number of hydrogen-bonding acceptor (NHBA) is important index for the hole mobility of organic semiconductors, seven dithienobenzothiazole (DBT) derivatives 1 a-g (NHBA=5) were designed and synthesized by one-step functionalization from a common precursor. X-ray single-crystal structural analyses confirmed that the molecular arrangements of 1b (the diethyl and ethylthienyl derivative) and 1c (the di(n-propyl) and n-propylthienyl derivative) in the crystal are classified into brickwork structures with multidirectional intermolecular charge-transfer integrals, as a result of incorporation of multiple hydrogen-bond acceptors. The solution-processed top-gate bottom-contact devices of 1b and 1c had hole mobilities of 0.16 and 0.029 cm2 V-1s-1, respectively.

Tackling Data Scarcity Challenge through Active Learning in Materials Processing with Electrospray

Machine learning (ML) has been harnessed as a promising modelling tool for materials research. However, small data, or data scarcity, is a bottleneck when incorporating ML in studies involving experimentation. Current experiment planning methods show several disadvantages: one‐factor‐at‐a‐time (OFAT) experimentation became impractical due to limited laboratory resources; conventional design of experiments (DoE) failed to incorporate high‐dimensional features in ML; Surrogate‐based or Bayesian optimization (BO) shifted the goal to optimize material properties rather than guiding training data accumulation. The present research proposes leveraging active learning (AL) to strategically select critical data for experimentation. Two AL strategies, query‐by‐Committee (QBC) algorithm and Greedy method, are benchmarked against random query baseline on various materials datasets. AL is shown to efficiently reduce model prediction errors with minimal additional experiment data. Investigation of hyperparameters revealed benefits of applying AL at an early stage of experimental dataset construction. Moreover, AL is implemented and validated for an in‐house materials development task ‐ electrospray modelling. AL exploration as a paradigm is highlighted to guide experiment design for efficient data accumulation purposes, and its potential for further ML modelling. In doing so, the power of ML is expected to be fully unleashed to experimental researchers.

Machine learning system to assess rice crop change detection from satellite-derived RGVI due to tropical cyclones using remote sensing dataset

Tropical cyclones strike massive devastation of agricultural crops in South Asian countries practically every year. For assessing impacts on primary land use and land cover (LULC) classes of Patuakhali district in Bangladesh, this study utilized the Landsat-8 image dataset to compute change detection indices of a consistent cyclone landfall period in the late rainy season (October and November). Different stage of cyclones based on their wind speed (km/h) were also analyzed to explore agricultural crop change detection due to the tropical cyclone (TC) Bulbul (made landfall on November 9th, 2019). Three years later, cyclone Sitrang made landfall on October 24th, 2022, and was employed for the test results that the Bulbul's rice growth vegetation index (RGVI)-CD trained in the study's machine learning (ML) system. Therefore, the objective of this study was to determine rice crop changes from RGVI due to a cyclone (Bulbul) to evaluate growth change status by another cyclone (Sitrang) using machine learning system. In Patuakhali district, the rice crops were damaged mainly and affected in classification as moderately changed detected (MCD) 26.07%, very changed detected (VCD) 48.83%, and extremely changed detected (ECD) 17.73% in total agricultural lands area (1548.93 km2). It was also significantly monitored for the rice crops to change as rice growth vegetation index (RGVI)-CD related to cyclone wind speed (km/h), DEM, distance from the rivers, and shorelines. From the result, the symmetric regression of coefficient (R2) of (i) linear regression, (ii) fine tree, (iii) support vector machine, (iv) ensemble boosted trees, and (v) trilayered neural network were obtained 0.735, 0.734, 0.737, 0.756 and 0.736 as good-fitted test outcomes respectively. Applying the five machine learning models that were created from cyclone Sitrang, it was discovered that the test set in this system was accurate to the observed forecast with acceptable RMSE and MAE values. Therefore, study can be helpful for researchers creating new damaged and yield loss area assessment policies to help TC-affected farming communities, especially rice growers, prepare an effective emergency response plan for their socioeconomic needs fulfillment in the changing climate perspectives for attaining sustainable development.

Quantum-Assisted Machine Learning for Enhanced Fraud Detection in Cybersecurity

Innovative methods for information security and fraud prevention are required in today's digital environment due to the expanding volume of data and the increasing complexity of cyber threats. The use of quantum computing techniques to improve fraud detection and classification systems is investigated in this study. The study's machine learning framework integrates three distinct quantum algorithms to improve classification techniques. The first technique uses a Pauli feature map and a Quantum Support Vector Classifier (QSVC) that leverages a quantum kernel to transform classical input into quantum states. The second technique use a ZZ feature map with "linear" entanglement and a support vector classifier model, utilizing quantum kernels to enhance quantum systems. The third method utilizes Variational Quantum Circuits (VQC) with actual amplitudes, which integrate quantum and conventional machine learning techniques to provide optimized classification. The best results were obtained by the QSVC using ZZ feature maps and linear entanglement, which had a precision of 1.0 and a notable decrease in false positives. In order to improve fraud detection systems' accuracy and dependability and offer strong solutions to financial institutions, this study shows how quantum computing has the potential to completely transform cybersecurity.

Predicting COVID-19 Mortality and Identifying Clinical Symptom Patterns in Hospitalized Patients: A Machine-learning Study

Background and Purpose: Identifying effective symptoms, demographic information, and underlying diseases to predict COVID-19 mortality is essential. We aimed to study the effective clinical and symptomatic characteristics of COVID-19 mortality in hospitalized patients with positive polymerase chain reaction (PCR) test results. Materials and Methods: For this study, we prospectively collected complete data on 26867 hospitalized individuals who tested PCR positive for COVID-19 from February 20, 2020, to September 12, 2021, in the Khorasan Razavi Province, Iran. We analyzed the data using artificial neural networks (ANN) and logistic regression (LR) models. Results: The accuracy of the ANN model was higher than the LR (90.27% versus 90.15%). The ten most important predictors that contributed to the prediction of death were decreasing consciousness level, cough, PO2 level, age, chronic kidney disease, fever, headache, smoking status, chronic blood diseases, and diarrhea using the ANN model. Conclusion: In conclusion, individuals suffering from chronic diseases such as cancer, kidney and blood diseases, as well as immunodeficiency are at a higher risk of mortality. This important finding can help decision-makers and medical professionals in their efforts to consider these conditions and provide effective preventative measures to reduce the risk of death.

Investigation of residual stresses of multi-layer multi-track components built by directed energy deposition: experimental, numerical, and time-series machine-learning studies

Investigation of residual stresses of multi-layer multi-track components built by directed energy deposition: experimental, numerical, and time-series machine-learning studies

Structural phase transition involving octahedron tilting and ion migration in metal-halide perovskites: A machine-learning study

Structural phase transition involving octahedron tilting and ion migration in metal-halide perovskites: A machine-learning study

Acoustic analysis in stuttering: a machine-learning study.

Stuttering is a childhood-onset neurodevelopmental disorder affecting speech fluency. The diagnosis and clinical management of stuttering is currently based on perceptual examination and clinical scales. Standardized techniques for acoustic analysis have prompted promising results for the objective assessment of dysfluency in people with stuttering (PWS). We assessed objectively and automatically voice in stuttering, through artificial intelligence (i.e., the support vector machine - SVM classifier). We also investigated the age-related changes affecting voice in stutterers, and verified the relevance of specific speech tasks for the objective and automatic assessment of stuttering. Fifty-three PWS (20 children, 33 younger adults) and 71 age-/gender-matched controls (31 children, 40 younger adults) were recruited. Clinical data were assessed through clinical scales. The voluntary and sustained emission of a vowel and two sentences were recorded through smartphones. Audio samples were analyzed using a dedicated machine-learning algorithm, the SVM to compare PWS and controls, both children and younger adults. The receiver operating characteristic (ROC) curves were calculated for a description of the accuracy, for all comparisons. The likelihood ratio (LR), was calculated for each PWS during all speech tasks, for clinical-instrumental correlations, by using an artificial neural network (ANN). Acoustic analysis based on machine-learning algorithm objectively and automatically discriminated between the overall cohort of PWS and controls with high accuracy (88%). Also, physiologic ageing crucially influenced stuttering as demonstrated by the high accuracy (92%) of machine-learning analysis when classifying children and younger adults PWS. The diagnostic accuracies achieved by machine-learning analysis were comparable for each speech task. The significant clinical-instrumental correlations between LRs and clinical scales supported the biological plausibility of our findings. Acoustic analysis based on artificial intelligence (SVM) represents a reliable tool for the objective and automatic recognition of stuttering and its relationship with physiologic ageing. The accuracy of the automatic classification is high and independent of the speech task. Machine-learning analysis would help clinicians in the objective diagnosis and clinical management of stuttering. The digital collection of audio samples here achieved through smartphones would promote the future application of the technique in a telemedicine context (home environment).

Dead or alive: Distinguishing active from passive particles using supervised learning (a)

A longstanding open question in the field of dense disordered matter is how precisely structure and dynamics are related to each other. With the advent of machine learning, it has become possible to agnostically predict the dynamic propensity of a particle in a dense liquid based on its local structural environment. Thus far, however, these machine-learning studies have focused almost exclusively on simple liquids composed of passive particles. Here we consider a mixture of both passive and active (i.e., self-propelled) Brownian particles, with the aim to identify the active particles from minimal local structural information. We compare a state-of-the-art machine learning approach for passive systems with a new method we develop based on Voronoi tessellation. Both methods accurately identify the active particles based on their structural properties at high activity and low concentrations of active particles. Our Voronoi method is, however, substantially faster to train and deploy because it requires fewer, and easy to compute, input features. Notably, both become ineffective when the activity is low, suggesting a fundamentally different structural signature for dynamic propensity and non-equilibrium activity. Ultimately, these efforts might also find relevance in the context of biological active glasses such as confluent cell layers, where subtle changes in the microstructure can hint at pathological changes in cell dynamics.

Efficient Screening of Metal Promoters of Pt Catalysts for C-H Bond Activation in Propane Dehydrogenation from a Combined First-Principles Calculations and Machine-Learning Study.

Platinum-based materials are the most widely used catalysts in propane direct dehydrogenation, which could achieve a balanced activity between both propane conversion and propene formation. One of the core issues of Pt catalysts is how to efficiently activate the strong C-H bond. It has been suggested that adding second metal promoters could greatly solve this problem. In the current work, first-principles calculations combined with machine learning are performed in order to obtain the most promising metal promoters and identify key descriptors for control performance. The combination of three different modes of adding metal promoters and two ratios between promoters and platinum sufficiently describes the system under investigation. The activity of propane activation and the formation of propene are reflected by the increase or decrease of the adsorption energy and C-H bond activation of propane and propene after the addition of promoters. The data of adsorption energy and kinetic barriers from first-principles calculations are streamed into five machine-learning methods including gradient boosting regressor (GBR), K neighbors regressor (KNR), random forest regressor (RFR), and AdaBoost regressor (ABR) together with the sure independence screening and sparsifying operator (SISSO). The metrics (RMSE and R2) from different methods indicated that GBR and SISSO have the most optimal performance. Furthermore, it is found that some descriptors derived from the intrinsic properties of metal promoters can determine their properties. In the end, Pt3Mo is identified as the most active catalyst. The present work not only provides a solid foundation for optimizing Pt catalysts but also provides a clear roadmap to screen metal alloy catalysts.

Spatial or Random Cross-Validation? The Effect of Resampling Methods in Predicting Groundwater Salinity with Machine Learning in Mediterranean Region

Machine learning (ML) algorithms are extensively used with outstanding prediction accuracy. However, in some cases, their overfitting capabilities, along with inadvertent biases, might produce overly optimistic results. Spatial data are a special kind of data that could introduce biases to ML due to their intrinsic spatial autocorrelation. To address this issue, a special resampling method has emerged called spatial cross-validation (SCV). The purpose of this study was to evaluate the performance of SCV compared with conventional random cross-validation (CCV) used in most ML studies. Multiple ML models were created with CCV and SCV to predict groundwater electrical conductivity (EC) with data (A) from Rhodope, Greece, in the summer of 2020; (B) from the same area but at a different time (summer 2019); and (C) from a new area (the Salento peninsula, Italy). The results showed that the SCV provides ML models with superior generalization capabilities and, hence, better prediction results in new unknown data. The SCV seems to be able to capture the spatial patterns in the data while also reducing the over-optimism bias that is often associated with CCV methods. Based on the results, SCV could be applied with ML in studies that use spatial data.

Comprehensive studies on the universality of BKT transitions—machine-learning study, Monte Carlo simulation, and level-spectroscopy method

Comprehensive studies are made on the six-state clock universality of two models using several approaches. We apply the machine-learning technique of phase classification to the antiferromagnetic (AF) three-state Potts model on the square lattice with ferromagnetic next-nearest-neighbor (NNN) coupling and the triangular AF Ising model with anisotropic NNN coupling to study two Berezinskii–Kosterlitz–Thouless transitions. We also use the Monte Carlo simulation paying attention to the ratio of correlation functions of different distances for these two models. The obtained results are compared with those of the previous studies using the level-spectroscopy method. We directly show the six-state clock universality for totally different systems with the machine-learning study.

Extracting Features of Active Transition Metal Electrodes for NO Electroreduction with Catalytic Matrices.

Electrocatalytic reduction of oxidized nitrogen compounds (NOx) promises to help rebalance the nitrogen cycle. It is widely accepted that nitrate reduction to NH4+/NH3 involves NO as an intermediate, and NO hydrogenation is the potential-limiting step of NO reduction. Whether *NO hydrogenates to *NHO or *NOH is still a matter of debate, which makes it difficult to optimize catalysts for NOx electroreduction. Here, "catalytic matrices" are used to swiftly extract features of active transition metal catalysts for NO electroreduction. The matrices show that active catalysts statistically stabilize *NHO over *NOH and have undercoordinated sites. Besides, square-symmetry active sites with Cu and other elements may prove active for NO electroreduction. Finally, multivariate regressions are able to reproduce the main features found by the matrices, which opens the door for more sophisticated machine-learning studies. In sum, catalytic matrices may ease the analysis of complex electrocatalytic reactions on multifaceted materials.

Automatic clustering and feature selection using multi-objective crow search algorithm

Today’s real-world data is frequently significant in size, with many redundant, missing, and noise-based features and data instances must be addressed before applying various data-mining-based algorithms for further knowledge discovery. Excessive dimensionality may be mitigated by carefully excluding unnecessary characteristics and selecting a reasonable subset of features. When presented as an optimization issue, choosing the best clusters using the most suitable subset of attributes is a challenge that may be handled using practical meta-heuristic approaches. Besides this, the automatic finding of the appropriate cluster number is another challenging task for the real-world dataset in the unsupervised machine-learning study. The present work proposes a multi-objective crow search algorithm for clustering and feature selection (MO-CSACFS) by modifying the crow search algorithm and introducing a levy flight-based two-point cross-over mechanism for a better exploration phase of the crow and further making it suitable for multi-objective optimization problems. MO-CSACFS addresses both issues using the three objective functions to find appropriate cluster numbers and features. MO-CSACFS is implemented over several real-life and synthetic datasets with varying instances, features, and cluster numbers to assess the algorithm’s performance; apart from that, the present work is also applied over several gene-expression datasets. MO-CSACFS is compared with two similar recently proposed multi-objective optimization processes used over an automatic, unsupervised machine learning task. The results show that the MO-CSACFS has produced a compact and robust cluster comparable to other similar works from the literature.

Developing a Warning Model of Potentially Inappropriate Medications in Older Chinese Outpatients in Tertiary Hospitals: A Machine-Learning Study.

Due to multiple comorbid illnesses, polypharmacy, and age-related changes in pharmacokinetics and pharmacodynamics in older adults, the prevalence of potentially inappropriate medications (PIMs) is high, which affects the quality of life of older adults. Building an effective warning model is necessary for the early identification of PIMs to prevent harm caused by medication in geriatric patients. The purpose of this study was to develop a machine learning-based model for the warning of PIMs in older Chinese outpatients. This retrospective study was conducted among geriatric outpatients in nine tertiary hospitals in Chengdu from January 2018 to December 2018. The Beers criteria 2019 were used to assess PIMs in geriatric outpatients. Three problem transformation methods were used to tackle the multilabel classification problem in prescriptions. After the division of patient prescriptions into the training and test sets (8:2), we adopted six widely used classification algorithms to conduct the classification task and assessed the discriminative performance by the accuracy, precision, recall, F1 scores, subset accuracy (ss Acc), and Hamming loss (hm) of each model. The results showed that among 11,741 older patient prescriptions, 5816 PIMs were identified in 4038 (34.39%) patient prescriptions. A total of 41 types of PIMs were identified in these prescriptions. The three-problem transformation methods included label power set (LP), classifier chains (CC), and binary relevance (BR). Six classification algorithms were used to establish the warning models, including Random Forest (RF), Light Gradient Boosting Machine (LightGBM), eXtreme Gradient Boosting (XGBoost), CatBoost, Deep Forest (DF), and TabNet. The CC + CatBoost model had the highest accuracy value (97.83%), recall value (89.34%), F1 value (90.69%), and ss Acc value (97.79%) with a good precision value (92.18%) and the lowest hm value (0.0006). Therefore, the CC + CatBoost model was selected to predict the occurrence of PIM in geriatric Chinese patients. This study's novelty establishes a warning model for PIMs in geriatric patients by using machine learning. With the popularity of electronic patient record systems, sophisticated computer algorithms can be implemented at the bedside to improve medication use safety in geriatric patients in the future.

Machine Learning in Tremor Analysis: Critique and Directions.

Tremor is the most frequent human movement disorder, and its diagnosis is based on clinical assessment. Yet finding the accurate clinical diagnosis is not always straightforward. Fine-tuning of clinical diagnostic criteria over the past few decades, as well as device-based qualitative analysis, has resulted in incremental improvements to diagnostic accuracy. Accelerometric assessments are commonplace, enabling clinicians to capture high-resolution oscillatory properties of tremor, which recently have been the focus of various machine-learning (ML) studies. In this context, the application of ML models to accelerometric recordings provides the potential for less-biased classification and quantification of tremor disorders. However, if implemented incorrectly, ML can result in spurious or nongeneralizable results and misguided conclusions. This work summarizes and highlights recent developments in ML tools for tremor research, with a focus on supervised ML. We aim to highlight the opportunities and limitations of such approaches and provide future directions while simultaneously guiding the reader through the process of applying ML to analyze tremor data. We identify the need for the movement disorder community to take a more proactive role in the application of these novel analytical technologies, which so far have been predominantly pursued by the engineering and data analysis field. Ultimately, big-data approaches offer the possibility to identify generalizable patterns but warrant meaningful translation into clinical practice. © 2023 The Authors. Movement Disorders published by Wiley Periodicals LLC on behalf of International Parkinson and Movement Disorder Society.

Machine-learning analysis of opioid use disorder informed by MOR, DOR, KOR, NOR and ZOR-based interactome networks

Opioid use disorder (OUD) continuously poses major public health challenges and social implications worldwide with dramatic rise of opioid dependence leading to potential abuse. Despite that a few pharmacological agents have been approved for OUD treatment, the efficacy of said agents for OUD requires further improvement in order to provide safer and more effective pharmacological and psychosocial treatments. Proteins including mu, delta, kappa, nociceptin, and zeta opioid receptors are the direct targets of opioids and play critical roles in therapeutic treatments. The protein–protein interaction (PPI) networks of the these receptors increase the complexity in the drug development process for an effective opioid addiction treatment. The report below presents a PPI-network informed machine-learning study of OUD. We have examined more than 500 proteins in the five opioid receptor networks and subsequently collected 74 inhibitor datasets. Machine learning models were constructed by pairing gradient boosting decision tree (GBDT) algorithm with two advanced natural language processing (NLP)-based autoencoder and Transformer fingerprints for molecules. With these models, we systematically carried out evaluations of screening and repurposing potential of more than 120,000 drug candidates for four opioid receptors. In addition, absorption, distribution, metabolism, excretion, and toxicity (ADMET) properties were also considered in the screening of potential drug candidates. Our machine-learning tools determined a few inhibitor compounds with desired potency and ADMET properties for nociceptin opioid receptors. Our approach offers a valuable and promising tool for the pharmacological development of OUD treatments.