Evaluate Machine Learning Algorithms Research Articles

A Survey of Learning Criteria Going beyond the Usual Risk (Abstract Reprint)

Virtually all machine learning tasks are characterized using some form of loss function, and "good performance" is typically stated in terms of a sufficiently small average loss, taken over the random draw of test data. While optimizing for performance on average is intuitive, convenient to analyze in theory, and easy to implement in practice, such a choice brings about trade-offs. In this work, we survey and introduce a wide variety of non-traditional criteria used to design and evaluate machine learning algorithms, place the classical paradigm within the proper historical context, and propose a view of learning problems which emphasizes the question of "what makes for a desirable loss distribution?" in place of tacit use of the expected loss.

Enhancing HLA-B27 antigen detection: Leveraging machine learning algorithms for flow cytometric analysis.

As the association of human leukocyte antigen B27 (HLA-B27) with spondylarthropathies is widely known, HLA-B27 antigen expression is frequently identified using flow cytometric or other techniques. Because of the possibility of cross-reaction with off target antigens, such as HLA-B7, each flow cytometric technique applies a "gray zone" reserved for equivocal findings. Our aim was to use machine learning (ML) methods to classify such equivocal data as positive or negative. Equivocal samples (n = 99) were selected from samples submitted to our institution for clinical evaluation by HLA-B27 antigen testing. Samples were analyzed by flow cytometry and polymerase chain reaction. Features of histograms generated by flow cytometry were used to train and validate ML methods for classification as logistic regression (LR), decision tree (DT), random forest (RF) and light gradient boost method (GBM). All evaluated ML algorithms performed well, with high accuracy, sensitivity, specificity, as well as negative and positive predictive values. Although, gradient boost approaches are proposed as high performance methods; nevertheless, their effectiveness may be lower for smaller sample sizes. On our relatively smaller sample set, the random forest algorithm performed best (AUC: 0.92), but there was no statistically significant difference between the ML algorithms used. AUC values for light GBM, DT, and LR were 0.88, 0.89, 0.89, respectively. Implementing these algorithms into the process of HLA-B27 testing can reduce the number of uncertain, false negative or false positive cases, especially in laboratories where no genetic testing is available.

Prediction of the Compressive Strength of Vibrocentrifuged Concrete Using Machine Learning Methods

The determination of mechanical properties for different building materials is a highly relevant and practical field of application for machine learning (ML) techniques within the construction sector. When working with vibrocentrifuged concrete products and structures, it is crucial to consider factors related to the impact of aggressive environments. Artificial intelligence methods can enhance the prediction of vibrocentrifuged concrete properties through the use of specialized machine learning algorithms for materials’ strength determination. The aim of this article is to establish and evaluate machine learning algorithms, specifically Linear Regression (LR), Support Vector Regression (SVR), Random Forest (RF), CatBoost (CB), for the prediction of compressive strength in vibrocentrifuged concrete under diverse aggressive operational conditions. This is achieved by utilizing a comprehensive database of experimental values obtained in laboratory settings. The following metrics were used to analyze the accuracy of the constructed regression models: Mean Absolute Error (MAE), Mean Squared Error (MSE), Root-Mean-Square Error (RMSE), Mean Absolute Percentage Error (MAPE) and coefficient of determination (R2). The average MAPE in the range from 2% (RF, CB) to 7% (LR, SVR) allowed us to draw conclusions about the possibility of using “smart” algorithms in the development of compositions and quality control of vibrocentrifuged concrete, which ultimately entails the improvement and acceleration of the construction and building materials manufacture. The best model, CatBoost, showed MAE = 0.89, MSE = 4.37, RMSE = 2.09, MAPE = 2% and R2 = 0.94.

Evaluating machine learning algorithms for prediction of response to ramelteon for sleep disturbances in patients with schizophrenia

Evaluating machine learning algorithms for prediction of response to ramelteon for sleep disturbances in patients with schizophrenia

Accuracy Comparison between Five Machine Learning Algorithms for Financial Risk Evaluation

An accurate prediction of loan default is crucial in credit risk evaluation. A slight deviation from true accuracy can often cause financial losses to lending institutes. This study describes the non-parametric approach that compares five different machine learning classifiers combined with a focus on sufficiently large datasets. It presents the findings on various standard performance measures such as accuracy, precision, recall and F1 scores in addition to Receiver Operating Curve-Area Under Curve (ROC-AUC). In this study, various data pre-processing techniques including normalization and standardization, imputation of missing values and the handling of imbalanced data using SMOTE will be discussed and implemented. Also, the study examines the use of hyper-parameters in various classifiers. During the model construction phase, various pipelines feed data to the five machine learning classifiers, and the performance results obtained from the five machine learning classifiers are based on sampling with SMOTE or hyper-parameters versus without SMOTE and hyper-parameters. Each classifier is compared to another in terms of accuracy during training and prediction phase based on out-of-sample data. The 2 data sets used for this experiment contain 1000 and 30,000 observations, respectively, of which the training/testing ratio is 80:20. The comparative results show that random forest outperforms the other four classifiers both in training and actual prediction.

Using machine learning algorithms with improved accuracy to analyze and predict employee attrition

Human migration is based on pull factors that individuals evaluate when it comes to moving to a different territory. Likewise, employee attrition is a phenomenon that represents the tendency to a reduction in employees within an organization. This research paper aims to develop and evaluate machine learning algorithms, namely Decision Tree, Random Forest, and Binary Logistic Regression, to predict employee attrition using the IBM dataset available on Kaggle. The objective is to provide organizations with a proactive approach to employee retention and human resource management by creating accurate predictive models. Employee attrition has significant implications for an organization's reputation, profitability, and overall structure. By accurately predicting employee attrition, organizations can identify the factors contributing to it and implement data-driven human resources management practices. This study contributes to improving decision-making processes, including hiring and firing decisions, and ultimately enhances an organization's capital. The IBM dataset used in this study consists of anonymized employee records and their employment outcomes. It provides a comprehensive HR data representation for analysis and prediction. Three machine learning algorithms, Decision Tree, Random Forest, and Binary Logistic Regression, were utilized in this research. These algorithms were selected for their potential to improve accuracy in predicting employee attrition. The Logistic Regression model yielded the highest accuracy of 87.44% among the tested algorithms. By leveraging this study's findings, organizations can develop predictive models to identify factors contributing to employee attrition. These insights can inform strategic decisions and optimize human resource management practices.

Use of Biochemical Tests and Machine Learning in the Search for Potential Diagnostic Biomarkers of COVID-19, HIV/AIDS, and Pulmonary Tuberculosis

This study aims to develop, validate, and evaluate machine learning algorithms for predicting the diagnosis of coronavirus disease (COVID-19), human immunodeficiency virus/acquired immunodeficiency syndrome (HIV/AIDS), pulmonary tuberculosis (TB), and HIV/TB co-infection. We also investigated potential biomarkers associated with the diagnosis. Data from biochemical and hematological tests of infected and controls were collected in a single general hospital, totalizing 6,418 patients. The discriminant analysis by partial least squares (PLS-DA) model had the highest performance in predicting the diagnosis of COVID-19, HIV/AIDS, TB, and HIV/TB co-infection with an accuracy of 94, 97, 95, and 96%, respectively. The biomarkers calcium, lactate dehydrogenase, red blood cells (RBC), white blood cells, neutrophils, basophils, eosinophils, hemoglobin, and hematocrit were associated with COVID-19. HIV infection was associated with mean corpuscular volume, platelets, neutrophils, and mean platelet volume. Red blood cell distribution width and urea were associated with infection by Mycobacterium tuberculosis. The following biomarkers were associated with HIV/TB co-infection: lymphocytes, RBC, hematocrit, hemoglobin, aspartate transaminase, alanine transaminase, and glycemia. The PLS-DA model can optimize COVID-19, HIV/AIDS, TB, and HIV/TB co-infection diagnostics. Some biomarkers were potential diagnostic indicators and could be evaluated during the screening of these diseases.

A Survey of Learning Criteria Going Beyond the Usual Risk

Virtually all machine learning tasks are characterized using some form of loss function, and “good performance” is typically stated in terms of a sufficiently small average loss, taken over the random draw of test data. While optimizing for performance on average is intuitive, convenient to analyze in theory, and easy to implement in practice, such a choice brings about trade-offs. In this work, we survey and introduce a wide variety of non-traditional criteria used to design and evaluate machine learning algorithms, place the classical paradigm within the proper historical context, and propose a view of learning problems which emphasizes the question of “what makes for a desirable loss distribution?” in place of tacit use of the expected loss.

利用人工智能技术改进高校课程评估与个性化学习

With continuous advancements in education and the rapid proliferation of artificial intelligence technology, leveraging AI to improve university course evaluation and personalize learning experiences has garnered significant interest. This article aims to investigate the effective utilization of artificial intelligence technology in enhancing university course evaluation and achieving personalized learning outcomes. Firstly, it provides an overview of the current landscape of university course evaluation and personalized learning while highlighting the limitations of existing methodologies. It then delves into the application of artificial intelligence technology to address these challenges, discussing the utilization of machine learning algorithms for course evaluation, personalized recommendations, and student data analysis. In conclusion, the article summarizes the potential impact of artificial intelligence technology on university course evaluation and personalized learning.

B-146 Discriminative Index: A Novel Indicator for Evaluating Machine Learning Algorithms in Laboratory Medicine

Abstract Background In clinical scenarios, incorrect predictions by machine learning (ML) models are inevitable. One way to reduce misleading predictions is to avoid reporting predictions that fall within a predefined “gray zone”. This method improves predictive performance by reporting only the less uncertain cases. However, the cost-effectiveness of applying the “gray zone” rule in an ML model is unclear without massive computation. Thus, this study aims to propose a novel metric to evaluate the effectiveness of using gray zones and validate the metric in real-world ML models. Methods This study defined a statistical metric called the “discriminative index” (D-index) for evaluating the effectiveness of gray zones. To calculate the D-index, the predictive outcomes of the ML model are first transformed into two probability distributions based on the truth labels (e.g., positive or negative). The D-index is then derived from the kurtosis of these two distributions. To validate the metric, we applied the D-index to three different antibiotic susceptibility-predicting ML models (namely, convolutional neural network (CNN), random forest (RF), and XGBoost (XGB)) based on mass spectrometry data. We assessed the performance and unpredicted case numbers of each model with different gray zones and correlated the results with the proposed D-index. Results The D-index values for the CNN, XGB, and RF models were 4.36, 0.38, and −1.66, respectively. When applying the “gray zone” rule to achieve 90% area under the receiver operating characteristic, the CNN, XGB and RF models retained up to 90%, 68%, and 62% of total cases, respectively. A higher D-index value indicates a more effective application of the gray zone rule. Conclusion The D-index is a simple and statistically insightful metric for evaluating the cost-effectiveness of applying gray zone rule in an ML model. This metric has been validated in three mass spectrometry-based predictive models and has shown promising results. The D-index can be a useful tool for comparing and applying different ML algorithms.

Evaluating machine learning algorithms for predicting compressive strength of concrete with mineral admixture using long short-term memory (LSTM) Technique

Evaluating machine learning algorithms for predicting compressive strength of concrete with mineral admixture using long short-term memory (LSTM) Technique

Improved patient mortality predictions in emergency departments with deep learning data-synthesis and ensemble models

The triage process in emergency departments (EDs) relies on the subjective assessment of medical practitioners, making it unreliable in certain aspects. There is a need for a more accurate and objective algorithm to determine the urgency of patients. This paper explores the application of advanced data-synthesis algorithms, machine learning (ML) algorithms, and ensemble models to predict patient mortality. Patients predicted to be at risk of mortality are in a highly critical condition, signifying an urgent need for immediate medical intervention. This paper aims to determine the most effective method for predicting mortality by enhancing the F1 score while maintaining high area under the receiver operating characteristic curve (AUC) score. This study used a dataset of 7325 patients who visited the Yonsei Severance Hospital’s ED, located in Seoul, South Korea. The patients were divided into two groups: patients who deceased in the ED and patients who didn’t. Various data-synthesis techniques, such as SMOTE, ADASYN, CTGAN, TVAE, CopulaGAN, and Gaussian Copula, were deployed to generate synthetic patient data. Twenty two ML models were then utilized, including tree-based algorithms like Decision tree, AdaBoost, LightGBM, CatBoost, XGBoost, NGBoost, TabNet, which are deep neural network algorithms, and statistical algorithms such as Support Vector Machine, Logistic Regression, Random Forest, k-nearest neighbors, and Gaussian Naive Bayes, as well as Ensemble Models which use the results from the ML models. Based on 21 patient information features used in the pandemic influenza triage algorithm (PITA), the models explained previously were applied to aim for the prediction of patient mortality. In evaluating ML algorithms using an imbalanced medical dataset, conventional metrics like accuracy scores or AUC can be misleading. This paper emphasizes the importance of using the F1 score as the primary performance measure, focusing on recall and specificity in detecting patient mortality. The highest-ranked model for predicting mortality utilized the Gaussian Copula data-synthesis technique and the CatBoost classifier, achieving an AUC of 0.9731 and an F1 score of 0.7059. These findings highlight the effectiveness of machine learning algorithms and data-synthesis techniques in improving the prediction performance of mortality in EDs.

Churn Identification and Prediction from a Large-Scale Telecommunication Dataset Using NLP

The identification of customer churn is a major issue for large telecom businesses. In order to manage the data of current customers as well as acquire and manage new customers, every day, a substantial volume of data gets generated. Therefore, it's crucial to identify the causes of client churn so that the appropriate steps can be taken to lower it. Numerous researchers have already discussed their efforts to combine static and dynamic approaches in order to reduce churn in big data sets, but these systems still have many issues when it comes to actually identifying churn. In this paper, we suggested two methods, the first of which is churn identification and using Natural Language Processing (NLP) methods and machine learning techniques, we make predictions based on a vast telecommunication data set. The NLP process involves data pre-processing, normalization, feature extraction, and feature selection. For feature extraction, we employ unique techniques like TF-IDF, Stanford NLP, and occurrence correlation methods, have been suggested. Throughout the lesson, a machine learning classification algorithm is used for training and testing. Finally, the system employs a variety of cross validation techniques and training and evaluating Machine learning algorithms. The experimental analysis shows the system's efficacy and accuracy.

Prediction of combustion, performance, and emission parameters of ethanol powered spark ignition engine using ensemble Least Squares boosting machine learning algorithms

This research concentrates on the application of machine learning techniques to predict combustion, performance, and emission parameters in a dual-fuel spark ignition (SI) engine powered by neat gasoline and E20 ethanol dual fuel. The goal is to overcome the limitations posed by repeated engine experiments and nonlinear test results. In order to optimise engine parameters, the research seeks to develop efficient machine learning models with high generalizability and employ an optimisation strategy to determine the optimal engine settings. Input for training and evaluating machine learning algorithms, such as Artificial Neural Networks (ANN), and Ensemble LS Boosting was derived from experimental data from a combustion test engine, which includes Neat gasoline, and Ethanol dual fuel blend E20 at various load conditions. The dataset includes engine combustion, performance, and emission indices such as brake thermal efficiency (BTE), Exhaust Gas Temperature (EGT), Hydrocarbons (HC), Carbon monoxide (CO), Carbon dioxide (CO2), and Nitrogen oxides (NOx), under various operating conditions. Load and brake-specific fuel consumption (BSFC) were training input attributes. Using a comprehensive experimental database of input-output engine parameters, the Artificial Neural Network (ANN) and Ensemble LS Boosting were constructed. The training data points were resampled to generate multiple training datasets for training different models. 50 test samples were used to evaluate the generalisation capability of the machine learning models, while BTE, EGT, CO, CO2, HC and NOx, were the primary parameters subject to prediction. The optimal machine learning method was determined by comparing R-squared (R2) values, root mean square error (RMSE), mean square error (MSE), and mean absolute error (MAE). Using multiple hyperparameter tuning iterations, the agreement between actual and predicted values for diverse Ensemble LS Boost algorithms was evaluated. The Ensemble LS Boost model exhibits the maximum level of agreement between predicted and experimental engine parameters across all datasets when compared to the other ANN models. This finding was corroborated by additional research based on test datasets, specifically the test sample interpolation data, which measures generalisation ability. The study also focuses on developing and applying two unique, interactive Simulink models for the Spark Ignition (SI) engine that are tailored for Neat Gasoline and Ethanol E20 test fuels under all loads. The key component of the model-based development technique in MATLAB and Simulink was the incorporation of sophisticated machine learning algorithms, i.e., Ensemble Least-Squares (LS) Boosting, to the model-based development workflow which produced reliable results. Implementing an Ensemble LS Boost machine learning framework is therefore highly recommended as an efficient method for predicting and optimising the combustion, performance, and emission characteristics of dual-fuel gasoline engines utilising Ethanol-based dual-fuel blends.

FracAtlas: A Dataset for Fracture Classification, Localization and Segmentation of Musculoskeletal Radiographs

Digital radiography is one of the most common and cost-effective standards for the diagnosis of bone fractures. For such diagnoses expert intervention is required which is time-consuming and demands rigorous training. With the recent growth of computer vision algorithms, there is a surge of interest in computer-aided diagnosis. The development of algorithms demands large datasets with proper annotations. Existing X-Ray datasets are either small or lack proper annotation, which hinders the development of machine-learning algorithms and evaluation of the relative performance of algorithms for classification, localization, and segmentation. We present FracAtlas, a new dataset of X-Ray scans curated from the images collected from 3 major hospitals in Bangladesh. Our dataset includes 4,083 images that have been manually annotated for bone fracture classification, localization, and segmentation with the help of 2 expert radiologists and an orthopedist using the open-source labeling platform, makesense.ai. There are 717 images with 922 instances of fractures. Each of the fracture instances has its own mask and bounding box, whereas the scans also have global labels for classification tasks. We believe the dataset will be a valuable resource for researchers interested in developing and evaluating machine learning algorithms for bone fracture diagnosis.

Evaluating machine learning algorithms to Predict 30-day Unplanned REadmission (PURE) in Urology patients

BackgroundUnplanned hospital readmissions are serious medical adverse events, stressful to patients, and expensive for hospitals. This study aims to develop a probability calculator to predict unplanned readmissions (PURE) within 30-days after discharge from the department of Urology, and evaluate the respective diagnostic performance characteristics of the PURE probability calculator developed with machine learning (ML) algorithms comparing regression versus classification algorithms.MethodsEight ML models (i.e. logistic regression, LASSO regression, RIDGE regression, decision tree, bagged trees, boosted trees, XGBoost trees, RandomForest) were trained on 5.323 unique patients with 52 different features, and evaluated on diagnostic performance of PURE within 30 days of discharge from the department of Urology.ResultsOur main findings were that performances from classification to regression algorithms had good AUC scores (0.62–0.82), and classification algorithms showed a stronger overall performance as compared to models trained with regression algorithms. Tuning the best model, XGBoost, resulted in an accuracy of 0.83, sensitivity of 0.86, specificity of 0.57, AUC of 0.81, PPV of 0.95, and a NPV of 0.31.ConclusionsClassification models showed stronger performance than regression models with reliable prediction for patients with high probability of readmission, and should be considered as first choice. The tuned XGBoost model shows performance that indicates safe clinical appliance for discharge management in order to prevent an unplanned readmission at the department of Urology.

Machine learning algorithms and computational validation of single-nucleotide polymorphisms of antioxidant enzymes and oxidative stress markers in neonates.

Aim: To evaluate machine learning algorithms (MLAs) for predicting factors (oxidative stress biomarkers [OSBs] and single-nucleotide polymorphism of the antioxidant enzymes)for respiratory distress syndrome (RDS) and significant alterations in the liver functions (SALVs). Materials & methods: MLAs were applied for predicting the RDS and SALV (with OSB and single-nucleotide polymorphisms in the antioxidant enzymes) with area under the curve(AUC)as the accuracy measure. Results: The C5.0 algorithm best predicted SALV (AUC: 0.63) with catalase as the most important predictor. Bayesian network best predicted RDS (AUC: 0.6) andENOS1 was the most important predictor. Conclusion: MLAs carry great potential in identifying the potential genetic and OSBsin neonatal RDS and SALV. Validation in prospective studies is needed urgently.

REVIEW PAPER ON PREDICTION OF WATER QUALITY PARAMETER USING MACHINE LEARNING

Surface water pollution become a nuisance for humankind as river water fulfill requirement of a major population and traditional method of water quality assessment and evaluation is inadequate in this era. So using advance method of machine learning in prediction of surface water proves to be helpful to prevent future water accident. As we seen many recent studies of water quality prediction and river water assessment using machine learning approach for better accuracy and less labor and to optimize its overall results. It’s become essential to review the recent studies which used Machine Learning algorithms for prediction, analysis, evaluation and assessment of river water quality and different models used in these studies for different environmental conditions. Machine learning models are superior to handle such complex and non linear data such as water quality parameters with greater accuracy, reliability, cost-effectiveness and efficiency as considered as great tool for surface Water Quality monitoring, prediction, future projects and help lawmakers in policy. In this report we reviewed around 17 research papers which uses machine learning approach from different journal and concise it to covers the structure of study, datasets used, methodology analysis, models performance, environment susceptibility, comparative analysis and assessments of Machine Learning models progress in river water quality research. For better management and control of surface water quality and its treatment, this study will help in understanding and analyzing the studies reviewed in this paper and its future application. We can conclude that research on Water Quality prediction using Machine Learning model are inadequate in the context future vulnerability, observing increasing pollution in recent years we require more research in this field. Finally, this study provides breakthrough in Surface Water Engineering and Management to give a new direction to fore coming studies and fortified it scope also gives a comparative approach for its implementation in new studies.

Predicting Pain in People With Sickle Cell Disease in the Day Hospital Using the Commercial Wearable Apple Watch: Feasibility Study.

Sickle cell disease (SCD) is a genetic red blood cell disorder associated with severe complications including chronic anemia, stroke, and vaso-occlusive crises (VOCs). VOCs are unpredictable, difficult to treat, and the leading cause of hospitalization. Recent efforts have focused on the use of mobile health technology to develop algorithms to predict pain in people with sickle cell disease. Combining the data collection abilities of a consumer wearable, such as the Apple Watch, and machine learning techniques may help us better understand the pain experience and find trends to predict pain from VOCs. The aim of this study is to (1) determine the feasibility of using the Apple Watch to predict the pain scores in people with sickle cell disease admitted to the Duke University SCD Day Hospital, referred to as the Day Hospital, and (2) build and evaluate machine learning algorithms to predict the pain scores of VOCs with the Apple Watch. Following approval of the institutional review board, patients with sickle cell disease, older than 18 years, and admitted to Day Hospital for a VOC between July 2021 and September 2021 were approached to participate in the study. Participants were provided with an Apple Watch Series 3, which is to be worn for the duration of their visit. Data collected from the Apple Watch included heart rate, heart rate variability (calculated), and calories. Pain scores and vital signs were collected from the electronic medical record. Data were analyzed using 3 different machine learning models: multinomial logistic regression, gradient boosting, and random forest, and 2 null models, to assess the accuracy of pain scores. The evaluation metrics considered were accuracy (F1-score), area under the receiving operating characteristic curve, and root-mean-square error (RMSE). We enrolled 20 patients with sickle cell disease, all of whom identified as Black or African American and consisted of 12 (60%) females and 8 (40%) males. There were 14 individuals diagnosed with hemoglobin type SS (70%). The median age of the population was 35.5 (IQR 30-41) years. The median time each individual spent wearing the Apple Watch was 2 hours and 17 minutes and a total of 15,683 data points were collected across the population. All models outperformed the null models, and the best-performing model was the random forest model, which was able to predict the pain scores with an accuracy of 84.5%, and a RMSE of 0.84. The strong performance of the model in all metrics validates feasibility and the ability to use data collected from a noninvasive device, the Apple Watch, to predict the pain scores during VOCs. It is a novel and feasible approach and presents a low-cost method that could benefit clinicians and individuals with sickle cell disease in the treatment of VOCs.

Cerebrospinal fluid and serum proteomic profiles accurately distinguish neuroaxonal dystrophy from cervical vertebral compressive myelopathy in horses

Cervical vertebral compressive myelopathy (CVCM) and equine neuroaxonal dystrophy/degenerative myeloencephalopathy (eNAD/EDM) are leading causes of spinal ataxia in horses. The conditions can be difficult to differentiate, and there is currently no diagnostic modality that offers a definitive antemortem diagnosis. Evaluate novel proteomic techniques and machine learning algorithms to predict biomarkers that can aid in the antemortem diagnosis of noninfectious spinal ataxia in horses. Banked serum and cerebrospinal fluid (CSF) samples from necropsy-confirmed adult eNAD/EDM (n = 47) and CVCM (n = 25) horses and neurologically normal adult horses (n = 45). . A subset of serum and CSF samples from eNAD/EDM (n=5) and normal (n=5) horses was used to evaluate the proximity extension assay (PEA). All samples were assayed by PEA for 368 neurologically relevant proteins. Data were analyzed using machine learning algorithms to define potential diagnostic biomarkers. Of the 368 proteins, 84 were detected in CSF and 146 in serum. Eighteen of 84 proteins in CSF and 30/146 in serum were differentially abundant among the 3 groups, after correction for multiple testing. Modeling indicated that a 2-protein test using CSF had the highest accuracy for discriminating among all 3 groups. Cerebrospinal fluid R-spondin 1 (RSPO1) and neurofilament-light (NEFL), in parallel, predicted normal horses with an accuracy of 87.18%, CVCM with 84.62%, and eNAD/EDM with 73.5%. Cross-species platform. Uneven sample size. Proximity extension assay technology allows for rapid screening of equine biologic matrices for potential protein biomarkers. Machine learning analysis allows for unbiased selection of highly accurate biomarkers from high-dimensional data.