ROC AUC Research Articles

Machine Learning Techniques for Heart Disease Prediction Using a Multi-Algorithm Approach

This analysis explores the efficiency of machine learning systems for heart disease identification through a multi-algorithm approach. The main objective is to identify the best performing algorithm for accurate disease prediction, improving clinical decision making. Using criteria including accuracy, precision, recall, F1 score, and recall, the study assessed four algorithms: Random Forest (RF), Naïve Bayes (NB), Support Vector Machine (SVM), and Decision Tree (DT). The results show that Random Forest outperforms the others, achieving 86.23% precision, 93.76% recall, 89.84% F1 score, and 88.41% accuracy. Random Forest gets an AUC ROC result of 0.94, so Random Forest is considered a superior model in this scenario, especially because it has higher accuracy. The algorithms showed a strong balance between sensitivity and specificity. Decision Tree showed reasonable performance with a precision of 84.18% and a recall of 90.27%, while Naïve Bayes recorded a precision of 87.68% and a recall of 87.03%. SVM showed a precision of 87.40% and a recall of 84.78%, indicating some limitations in capturing positive cases. The novelty of this study lies in the comparative analysis of several algorithms to optimize the heart disease prediction model for clinical use. The random forest algorithm is one of the choices, but there is still a medical standard for classifying people as either indicating or not experiencing heart failure, according to the study.

Computerized tomography features acting as predictors for invasive therapy in the management of Crohn's disease-related spontaneous intra-abdominal abscess: experience from long-term follow-up.

Decision-making in the management of Crohn's disease (CD)-related spontaneous intra-abdominal abscess (IAA) is challenging. This study aims to reveal predictive factors for percutaneous drainage and/or surgery in the treatment of CD-related spontaneous IAA through long-term follow-up. Data were collected, including clinical manifestations, radiography and treatment strategies, in Chinese patients with CD-related IAA in a tertiary medical center. Univariate and Multivariate Cox analysis were conducted to identify predictors for invasive therapy. Altogether, 48 CD patients were identified as having IAA through enhanced CT scans. The median follow-up time was 45.0 (23.3, 58.0) months. 23 (47.9%) patients underwent conservative medical treatment, and 25 (52.1%) patients underwent percutaneous drainage and/or surgical intervention (invasive treatment group). The 1-, 2-, and 5-year overall survival rates without invasive treatment were 75.0%, 56.1%, and 46.1%, respectively. On univariate Cox analysis, the computerized tomography (CT) features including nonperienteric abscess (HR: 4.22, 95% CI: 1.81-9.86, p = 0.001), max abscess diameter (HR: 1.01, 95% CI: 1.00-1.02, p<0.001) and width of sinus (HR: 1.27, 95% CI: 1.10-1.46, p = 0.001) were significantly associated with invasive treatment. Nonperienteric abscess was significantly associated with invasive treatment on multivariate Cox analysis (HR: 3.11, 95% CI: 1.25-7.71, p = 0.015). A score model was built by width of sinus, location of abscess and max abscess diameter to predict invasive treatment. The AUC of ROC, sensitivity and specificity were 0.892, 80.0% and 90.9% respectively. More than half of CD-related IAA patients needed invasive therapy within 5-year follow-up. The CT features including nonperienteric abscess, larger maximum abscess diameter and width of sinus suggested a more aggressive approach to invasive treatment.

Predicting full-thickness necrosis in adult acute corrosive ingestion injuries in a sub-Saharan African setting

BACKGROUND Corrosive ingestion remains an important global pathology with high morbidity and mortality. Data on the acute management of adult corrosive injuries from sub-Saharan Africa is scarce, with international investigative algorithms, relying heavily on computed tomography (CT), having limited availability in this setting. AIM To investigate the corrosive injury spectrum in a low-resource setting and the applicability of parameters for predicting full-thickness (FT) necrosis and mortality. METHODS A retrospective analysis of a prospective corrosive injury registry (March 1, 2017–October 31, 2023) was performed to include all adult patients with acute corrosive ingestion managed at a single, academic referral centre in Cape Town, South Africa. Patient demographics, corrosive ingestion details, initial investigations, management, and short-term outcomes were described using descriptive statistics while multivariate analysis with receiver operator characteristic area under the curve graphs (ROC AUC) were used to identify factors predictive of FT necrosis and 30-day mortality. RESULTS One-hundred patients were included, with a mean age of 32 years (SD: 11.2 years) and a male predominance (65.0%). The majority (73.0%) were intentional suicide attempts. Endoscopy on admission was the most frequent initial investigation performed (95 patients), while only 17 were assessed with CT. Seventeen patients had full thickness necrosis at surgery, of which eleven underwent emergency resection and six were palliated. Thirty-day morbidity and mortality were 27.0% and 14.0%, respectively. Patients with full thickness necrosis and those with an established perforation had a 30-day mortality of 58.8% and 91.0%, respectively. Full thickness necrosis was associated with a cumulative 2-year survival of only 17.6%. Multivariate analyses with ROC AUC showed admission endoscopy findings, CT findings, and blood gas findings (pH, base excess, lactate), to all have significant predictive value for full thickness necrosis, with endoscopy proving to have the best predictive value (AUC 0.850). CT and endoscopy findings were the only factors predictive of early mortality, with CT performing better than endoscopy (AUC 0.798 vs 0.759). CONCLUSION Intentional corrosive injuries result in devastating morbidity and mortality. Locally, early endoscopy remains the mainstay of severity assessment, but referral for CT imaging should be considered especially when blood gas findings are abnormal.

Time Series Classification for Predicting Biped Robot Step Viability

The prediction of the stability of future steps taken by a biped robot is a very important task, since it allows the robot controller to adopt the necessary measures in order to minimize damages if a fall is predicted. We present a classifier to predict the viability of a given planned step taken by a biped robot, i.e., if it will be stable or unstable. The features of the classifier are extracted from a feature engineering process exploiting the useful information contained in the time series generated in the trajectory planning of the step. In order to state the problem as a supervised classification one, we need the ground truth class for each planned step. This is obtained using the Predicted Step Viability (PSV) criterion. We also present a procedure to obtain a balanced and challenging training/testing dataset of planned steps that contains many steps in the border between stable and non stable regions. Following this trajectory planning strategy for the creation of the dataset we are able to improve the robustness of the classifier. Results show that the classifier is able to obtain a 95% of ROC AUC for this demanding dataset using only four time series among all the signals required by PSV to check viability. This allows to replace the PSV stability criterion, which is safe, robust but impossible to apply in real-time, by a simple, fast and embeddable classifier that can run in real time consuming much less resources than the PSV.

Comparativa de diferentes modelos radiómicos para la clasificación de lesiones adrenales indeterminadas diagnosticadas de forma incidental en TC con contraste

Purpose: A comparison of different machine learning models to discriminate adrenal incidentalomas by CT studies was performed. Methods: Sixty-two different features were obtained from a sample of 61 incidentalomas using the free license software LIFEx and 19 radiomic studies were performed with different models and feature selection methods to obtain the most efficient determination of possible malignancy. For all of them, four cross-validation methods were evaluated. Adenoma contouring was performed in duplicate by different radiologists evaluating all models in both groups. Results: ROC AUC between 0.42 (0.09-0.81) and 0.92 (0.63-1.00), and accuracy of the models between 0.63 (0.43-0.79) and 0.94 (0.82-1.00). The best-performing model was the balanced logistic regression applied to the 14 features with an intraclass coefficient greater than 0.9, with which accuracy of 0.94 (0.74-1.00), ROC AUC of 0.917 (0.63-1.00), benign recall of 0.92 (0.65-1.00) and malignant recall of 1.00 (0.71-1.00) were obtained. Conclusions: The evaluation and validation of different models has allowed us to obtain an efficient radiomic model for the discrimination of adrenal incidentalomas

The impact of radiologists’ characteristics on the detection of COVID-19 in chest CT scans

PurposeTo retrospectively assess the relationships between radiologists’ characteristics and their performance in interpreting chest CTs for COVID-19 detection. MethodsUsing DetectedX software, performance data were collected online from 74 radiologists from 29 countries, who each interpreted 30 anonymized digital chest CTs (15 positive and 15 negative scans for COVID-19). Radiologist-specific information was also collected: radiology experience, radiological specialty, number of lung CTs read weekly, familiarity with and education about COVID-19 presentation, and suspected COVID-19 cases seen weekly. The influence of this radiologist-specific information on the radiologists’ sensitivity, specificity, and ROC AUC was determined using regression analysis. ResultsRadiologists without respiratory imaging specialization had greater ROC AUCs (0.83 vs. 0.70, p = .006) and sensitivities (74.0% vs. 47.7%, p = .002) than their specialized peers. Radiologists without COVID-19 case encounters had greater sensitivity (80.6% vs. 63.1%, p = .017) but lower specificity (71.1% vs. 83.4%, p = .014) than those who encountered at least one case weekly. ConclusionSpecialization and prior experience with COVID-19 may impact the interpretation of suspected COVID-19 chest CTs. Clinical significanceThese results show the importance of diverse expertise and continuous training in managing novel diseases like COVID-19. Incorporating varied perspectives and experiences could enhance diagnostic accuracy and ensure a more comprehensive approach to disease management. In the case of novel diseases, specialization may not provide the same advantage as with more familiar disorders. The diagnostic value of specific experience with a novel disorder may help to compensate for a lack of specialization in the field, particularly in emergent situations. While specialization is important, generalization also holds value and utility in appropriate contexts.

Skin graft take rate among minor burns – A cohort study to investigate the effect of factors such as burn depth, burn size, and timing of the operation

BackgroundIt has been known for decades that early excision of burns reduces morbidity and mortality. Early surgical excision and skin grafting has become the most important part of successful healing in burn management, especially in major burns. However it is not entirely clear whether early excision and skin transplantation has the same advantages in smaller burns and there is no consensus on the timing of skin grafting in this group. The aim was to investigate the effect of timing and other factors for skin graft take rate among minor burns. MethodsThis retrospective study included patients with thermal injury, a burn size smaller than 11 % total body surface area (TBSA), and who were treated with a skin graft operation. Take rate at the second dressing change after operation was used as main outcome, a cut-off of 95% take rate was for the multivariable logistic regression. ResultsA total of 195 patients were included, median (IQR) age was 42 (9–68) years, 65 % were male, and median (IQR) area of deep burns was 2 (1–4) % of the body surface area (BSA). Multivariable regression showed that smaller area of deep burns and scalds (compared with flame and contact burns) were associated with a take rate of ≥ 95 %. Age, timing of the skin graft transplantation, and plasma C-reactive protein showed no independent effect on take rate. The regression model was significant but weak (ROC AUC 0.71, 95 % CI 0.62–0.79). ConclusionOur results suggest that the extent and depth of the burn are the most important factors for skin graft take rate among minor burns, while timing of the transplantation is not associated with take rate for the skin graft. The advantageous effect of scalds may be interpreted to mean that scalds in general are more superficial than flame and contact burns, a difference that may not be detected by the use of a Lund and Browder chart. The conclusion is, however, tempered by the retrospective study design and the relatively low discriminatory power in our study.

External validation of a multimodality deep-learning normal tissue complication probability model for mandibular osteoradionecrosis trained on 3D radiation distribution maps and clinical variables

Background and purposeWhile the inclusion of spatial dose information in deep learning (DL)-based normal-tissue complication probability (NTCP) models has been the focus of recent research studies, external validation is still lacking. This study aimed to externally validate a DL-based NTCP model for mandibular osteoradionecrosis (ORN) trained on 3D radiation dose distribution maps and clinical variables. Methods and materialsA 3D DenseNet-40 convolutional neural network (3D-mDN40) was trained on clinical and radiation dose distribution maps on a retrospective class-balanced matched cohort of 184 subjects. A second model (3D-DN40) was trained on dose maps only and both DL models were compared to a logistic regression (LR) model trained on DVH metrics and clinical variables. All models were externally validated by means of their discriminative ability and calibration on an independent dataset of 82 subjects. ResultsNo significant difference in performance was observed between models. In internal validation, these exhibited similar Brier scores around 0.2, Log Loss values of 0.6–0.7 and ROC AUC values around 0.7 (internal) and 0.6 (external). Differences in clinical variable distributions and their effect sizes were observed between internal and external cohorts, such as smoking status (0.6 vs. 0.1) and chemotherapy (0.1 vs. −0.5), respectively. ConclusionTo our knowledge, this is the first study to externally validate a multimodality DL-based ORN NTCP model. Utilising mandible dose distribution maps, these models show promise for enhancing spatial risk assessment and guiding dental and oncological decision-making, though further research is essential to address overfitting and domain shift for reliable clinical use.

Bone Scintigraphy in Cardiac Transthyretin-Related Amyloidosis: A Novel Time-Saving Tool for Semiquantitative Analysis, with Good Potential for Predicting Different Etiologies

(1) Background: The visual and semiquantitative analysis of Technetium-99metastable-3,3-diphospono-1,2-propanodicarboxylic acid (99mTc-DPD) bone scintigraphy is promising for diagnosing cardiac amyloidosis but time-consuming. We validated a faster method, the geometric mean (GM) method with a semi-automated workflow, for heart–whole body (WB) ratio (H/WBr), heart retention (Hr), and WB retention (WBr) calculations compared to the classic method (CM) established in the literature. The capability of semiquantitative scintigraphy indexes to differentiate the etiology in transthyretin-related cardiac amyloidosis (cATTR) patients was investigated. (2) Methods: H/WBr, Hr, and WBr were calculated by extracting counts for WB, kidneys, bladder, and heart on early and late planar image scans and applying background, scan-time, and decay corrections, using CM and GM both on a referring workstation and on a semi-automated workflow in external software. The comparison between CM and GM was assessed with Pearson’s correlation, Lin’s Concordance Correlation Coefficient (CCC), and Bland–Altman analysis. H/WBr, Hr, and WBr and several clinical variables were used to implement LASSO, Random Forest (RF), and Neural Network (NN) models to predict mutated and wild-type ATTR etiologies. ROC curves and AUC were calculated. (3) Results: Hr, WBr, and H/WBr using CM and GM were highly correlated. Bland–Altman analysis between CM and GM showed biases of 0.12% [CI:0.04%;0.19%] for H/WBr, 0.07% [CI: 0.01%; 0.13%] for Hr, and -0.50% [CI: −1.22%; 0.22%] for WBr. LASSO and NN models had good performance in predicting etiologies with AUC values of 87.3% and 73.6%, respectively. The RF model showed a poorer AUC of 55.8%. (4) Conclusions: The GM in the assisted workflow was validated against the CM. LASSO and NN approaches allowed a good prediction performance to be obtained for patient etiology.

Deep Learning Ensemble for Flood Probability Analysis

Predicting flood events is complex due to uncertainties from limited gauge data, high data and computational demands of traditional physical models, and challenges in spatial and temporal scaling. This research innovatively uses only three remotely sensed and computed factors: rainfall, runoff and temperature. We also employ three deep learning models—Feedforward Neural Network (FNN), Convolutional Neural Network (CNN), and Long Short-Term Memory (LSTM)—along with a deep neural network ensemble (DNNE) using synthetic data to predict future flood probabilities, utilizing the Savitzky–Golay filter for smoothing. Using a hydrometeorological dataset from 1993–2022 for the Nile River basin, six flood predictors were derived. The FNN and LSTM models exhibited high accuracy and stable loss, indicating minimal overfitting, while the CNN showed slight overfitting. Performance metrics revealed that FNN achieved 99.63% accuracy and 0.999886 ROC AUC, CNN had 95.42% accuracy and 0.893218 ROC AUC, and LSTM excelled with 99.82% accuracy and 0.999967 ROC AUC. The DNNE outperformed individual models in reliability and consistency. Runoff and rainfall were the most influential predictors, while temperature had minimal impact.

Cox Regression Model with Time-varying Coefficients Applied to Survival Estimation of Heart Failure

This study focuses on survival analysis of heart failure patients. All the patients, who are over 40 years old, have NYHA classes III or IV stages of heart failure. Cox regression model with time-varying coefficients is applied to deal with these data, which is updated in 2015. Kaplan Meier curves are plotted to screen out categorical variables that violates PH assumptions. As for continuous variables violating PH assumption, Schoenfeld residual plot is drawn to realize time stratification. Therefore, for each time period, cox proportional hazard model can be fit to estimate mortality and find out the key factors promoting or inhibiting death event of heart failure. As a result, age, blood pressure, serum creatinine, and ejection fraction are the significant features of hazard of death led by heart failure, which is validated by Wald test, Likelihood ratio test, and Log-rank test. Discrimination ability is also characterized by ROC curves and AUC finally.

Enhancing Diagnostic Accuracy of Fresh Vertebral Compression Fractures with Deep Learning Models.

Retrospective study. The study aimed to develop and authenticated a deep learning model based on X-ray images to accurately diagnose fresh thoracolumbar vertebral compression fractures. In clinical practice, diagnosing fresh vertebral compression fractures often requires MRI. However, due to the scarcity of MRI resources and the high time and economic costs involved, some patients may not receive timely diagnosis and treatment. Using a deep learning model combined with X-rays for diagnostic assistance could potentially serve as an alternative to MRI. In this study, the main collection included X-ray images suspected of thoracolumbar vertebral compression fractures from the municipal shared database between December 2012 and February 2024. Deep learning models were constructed using frameworks of EfficientNet, MobileNet, and MnasNet respectively. We conducted a preliminary evaluation of the deep learning model using the validation set. The diagnostic performance of the models was evaluated using metrics such as AUC value, accuracy, sensitivity, specificity, F1 score, precision, and ROC curve. Finally, the deep learning models were compared with evaluations from two spine surgeons of different experience levels on the control set. This study included a total of 3025 lateral X-ray images from 2224 patients. The dataset was divided into a training set of 2388 cases, a validation set of 482 cases, and a control set of 155 cases. In the validation set, the three groups of DL models had accuracies of 83.0%, 82.4%, and 82.2%, respectively. The AUC values were 0.861, 0.852, and 0.865, respectively. In the control set, the accuracies of the three groups of DL models were 78.1%, 78.1%, and 80.7%, respectively, all higher than spinal surgeons and significantly higher than junior spine surgeon. This study developed deep learning models for detecting fresh vertebral compression fractures, demonstrating high accuracy.

Random Forest Prognostication of Survival and 6-Month Outcome In Pediatric Patients Following Decompressive Craniectomy For Traumatic Brain Injury

There is a dearth of literature regarding prognostic and predictive factors for outcome following pediatric decompressive craniectomy (DC) following traumatic brain injury (TBI). The aim of this study was to develop a random forest machine learning algorithm to predict outcomes following DC in pediatrics. This is a multi-institutional retrospective study assessing the 6-month postoperative outcome in pediatric patients that underwent DC. We developed a machine learning model using classification and survival random forest algorithms (CRF and SRF respectively) for the prediction of outcomes. Data was collected on clinical signs, radiographic studies, and laboratory studies. The outcome measures for the CRF were mortality and good or bad outcome based on Glasgow Outcome Scale (GOS) at 6-months. A GOS score of 4 or higher indicated a good outcome. The outcomes for the SRF model assessed mortality at during the follow-up period. A total of 40 pediatric patients were included. There was a hospital mortality rate of 27.5%, and 75.8% of survivors had a good outcome at 6-month follow up. The CRF for 6-month mortality had a ROC AUC of 0.984; whereas, the 6-month good/bad outcome had a ROC AUC of 0.873. The SRF was trained at the 6-month timepoint with a ROC AUC of 0.921. CRF and SRF models successfully predicted 6-month outcomes and mortality following DC in pediatric TBI patients. These results suggest random forest models may be efficacious for predicting outcome in this patient population.

Low conduction velocity is a key factor for localisation of reentrant drivers for atrial fibrillation

Abstract Background Atrial Fibrillation (AF) is a prevalent cardiac arrhythmia globally, associated with heightened risks of stroke, dementia, and heart failure. Catheter ablation, the sole curative therapy for AF, yields suboptimal success rates for persistent AF cases. Understanding the mechanisms and dynamics of AF remains challenging, hindering treatment advancements. Novel ablation strategies aimed at targeting substrates harbouring AF reentrant drivers (RD), such as low-voltage ablation and rotor-based ablation, have yet to demonstrate superiority over the gold standard ablation strategy – pulmonary vein isolation. Purpose The purpose of the study is to integrate in-silico 3D left atrial (LA) simulations and explainable artificial intelligence (AI) to pinpoint key factors for RD localisation in persistent AF, aiming to improve the mechanistic understanding of AF and help develop a more effective ablation strategy. Methods 3D LA models were derived from magnetic resonance images of 41 patients and combined with atrial cell electrophysiology models. Persistent AF was initiated in each patient-specific LA model with fast pacing, and five functional features (action potential duration (APD), conduction velocity (CV), wavelength, APD spatial gradient, CV spatial gradient, and wavelength spatial gradient) and three structural features (fibrotic density, fibrotic entropy, and image intensity ratio) were evaluated. A random forest AI model was trained to classify phase singularities corresponding to the RDs (PS-RD) in five distinct tissue classes (total LA tissue, healthy tissue, fibrotic border zone, dense fibrotic tissue, and PS-RD tissue) based on the functional and structural features. Shapley additive explanations (SHAP) analysis was then used to find the most influential feature in the AI model’s decision process. Results Simulations of the patient-specific LA models revealed that RDs were typically localised in regions with the lowest atrial CV, which often coincided with patches of dense fibrotic tissue. In a five-fold cross-validation, the random forest AI model achieved high accuracy in classifying PS-RD regions: a mean ROC AUC of 0.99 ± 0.010, a mean F1 score of 0.86 ± 0.034, a mean recall of 0.95 ± 0.058, and a mean precision score of 0.79 ± 0.085. The further SHAP analysis revealed that low CV was the key determinant in classifying the PS-RD regions. Conclusion Our study combined AF computational modelling and explainable AI to identify the key factors in RD localisation. Both the computational and AI models concluded that low CV was the most important determinant in localising RDs. These findings provide mechanistic evidence for the efficacy of ablation strategies that target low CV atrial areas for persistent AF patients.

Evaluation of Previous ECG Algorithms for Identifying the Culprit Artery in Inferior STEMI

Abstract Background ST elevation myocardial infarction (STEMI) is the leading cause of cardiovascular related mortality. In the context of Inferior STEMI, the culprit artery is whether the right coronary artery (RCA) or left circumflex coronary artery (LCx). An accurate prediction of culprit artery prior to primary percutaneous coronary intervention (PCI) could improve the door-to-balloon time and could prevent contrast-induced-nephropathy. However algorithms established for that purpose were not evaluated within the Tunisian population. We aimed then to assess their performance in our population Methods We conducted a single-center hostorical cohort study from January 2014 to December 2022 on patients admitted at Sahloul university hospital Sousse Tunisia for PCI following an inferior STEMI. We analyzed the ST segment deviations at the J-point and 80 ms after the J-point then evaluated the performance of previous algorithms. Results We collected data from 156 patients, of which 130 (83.3%) hada RCA occlusion, and 26 (16.7%) a LCx occlusion. We evaluated 19 published algorithms. The criterion of zimetbaum et al (ST elevation in III&amp;gt; II and I and/or VL &amp;lt;-1 mm) had the highest AUC 0.77[0.66 - 0.88] YI = 52 accuracy = 0.81. The criterion of Herz et al (ST III &amp;gt; II) had AUC ROC of 0.71 [0.58 - 0.83] YI = 0.41 accuracy = 0.78, the algorithm of Tierala et al: AUC = 0.70[0.58 - 0.83] YI = 0.40 accuracy = 0.78. The algorithm og Fiol et al: AUC = 0.79 [0.58-0.83] YI = 0.59 accuracy = 0.73. Conclusions The Zimetbaum criterion had the highest performance in predicting the culprit artery and Fiol’s algorithm was the most specific to the LCx occlusion in our population. Further study on the impact of these algorithms on patients outcome is necessary. Key messages • The prediction of the culprit artery in inferior STEMI is challenging. • The criterion of Zimetbaum et al had a good performance for that regard.

Innovative Sentiment Analysis and Prediction of Stock Price Using FinBERT, GPT-4 and Logistic Regression: A Data-Driven Approach

This study explores the comparative performance of cutting-edge AI models, i.e., Finaance Bidirectional Encoder representations from Transsformers (FinBERT), Generatice Pre-trained Transformer GPT-4, and Logistic Regression, for sentiment analysis and stock index prediction using financial news and the NGX All-Share Index data label. By leveraging advanced natural language processing models like GPT-4 and FinBERT, alongside a traditional machine learning model, Logistic Regression, we aim to classify market sentiment, generate sentiment scores, and predict market price movements. This research highlights global AI advancements in stock markets, showcasing how state-of-the-art language models can contribute to understanding complex financial data. The models were assessed using metrics such as accuracy, precision, recall, F1 score, and ROC AUC. Results indicate that Logistic Regression outperformed the more computationally intensive FinBERT and predefined approach of versatile GPT-4, with an accuracy of 81.83% and a ROC AUC of 89.76%. The GPT-4 predefined approach exhibited a lower accuracy of 54.19% but demonstrated strong potential in handling complex data. FinBERT, while offering more sophisticated analysis, was resource-demanding and yielded a moderate performance. Hyperparameter optimization using Optuna and cross-validation techniques ensured the robustness of the models. This study highlights the strengths and limitations of the practical applications of AI approaches in stock market prediction and presents Logistic Regression as the most efficient model for this task, with FinBERT and GPT-4 representing emerging tools with potential for future exploration and innovation in AI-driven financial analytics.

Bayesian intelligence for medical diagnosis: a pilot study on patient disposition for emergency medicine chest pain.

Clinicians can rapidly and accurately diagnose disease, learn from experience, and explain their reasoning. Computational Bayesian medical decision-making might replicate this expertise. This paper assesses a computer system for diagnosing cardiac chest pain in the emergency department (ED) that decides whether to admit or discharge a patient. The system can learn likelihood functions by counting data frequency. The computer compares patient and disease data profiles using likelihood. It calculates a Bayesian probabilistic diagnosis and explains its reasoning. A utility function applies the probabilistic diagnosis to produce a numerical BAYES score for making a medical decision. We conducted a pilot study to assess BAYES efficacy in ED chest pain patient disposition. Binary BAYES decisions eliminated patient observation. We compared BAYES to the HEART score. On 100 patients, BAYES reduced HEART's false positive rate 18-fold from 58.7 to 3.3 %, and improved ROC AUC accuracy from 0.928 to 1.0. The pilot study results were encouraging. The data-driven BAYES score approach could learn from frequency counting, make fast and accurate decisions, and explain its reasoning. The computer replicated these aspects of diagnostic expertise. More research is needed to reproduce and extend these finding to larger diverse patient populations.

Two-step GLIM approach using NRS-2002 screening tool vs direct GLIM criteria application in hospital malnutrition diagnosis: A cross-sectional study.

The two-step Global Leadership Initiative on Malnutrition (GLIM) approach was recently introduced to malnutrition diagnosis in a hospital setting. This study compares the diagnostic performance of this approach that uses the Nutritional Risk Screening-2002 (NRS-2002) as a screening tool and the direct application of GLIM malnutrition diagnostic criteria in hospitalized patients. This cross-sectional study involved 290 adult and older adult patients who were hospitalized. A trained nutritionist implemented the two-step GLIM approach, including NRS-2002 (in the first step) and GLIM criteria (in the second step) for each patient. Then, the accuracy, kappa index, area under the receiver operating characteristic curve (AUC ROC), sensitivity, and specificity of malnutrition diagnostic performance of the two-step GLIM approach and NRS-2002 were evaluated compared to thedirect use of GLIM criteria. The NRS-2002 identified 145 (50.0%) patients as at risk of malnutrition. The prevalence of malnutrition using the two-step GLIM approach and GLIM malnutrition diagnosis criteria were 120 (41.4%) and 141 (48.6%), respectively. The kappa index showed substantial and almost perfect agreement for NRS-2002 (κ = 68%) and the two-step GLIM approach (κ = 85%) with GLIM malnutrition diagnostic criteria, respectively. Furthermore, the AUC ROC (0.926; 95% confidence interval (CI): 0.89-0.96) and accuracy (92.8%) of the two-step GLIM approach compared to the GLIM criteria indicated an acepptable ability to distinguish between malnourished and well-nourished patients. The two-step GLIM approach using NRS-2002 as a screening tool (in step one) had acceptable malnutrition diagnostic performance compared to the direct application of GLIM criteria in hospitalized patients.

MEF-AlloSite: an accurate and robust Multimodel Ensemble Feature selection for the Allosteric Site identification model

A crucial mechanism for controlling the actions of proteins is allostery. Allosteric modulators have the potential to provide many benefits compared to orthosteric ligands, such as increased selectivity and saturability of their effect. The identification of new allosteric sites presents prospects for the creation of innovative medications and enhances our comprehension of fundamental biological mechanisms. Allosteric sites are increasingly found in different protein families through various techniques, such as machine learning applications, which opens up possibilities for creating completely novel medications with a diverse variety of chemical structures. Machine learning methods, such as PASSer, exhibit limited efficacy in accurately finding allosteric binding sites when relying solely on 3D structural information.Scientific ContributionPrior to conducting feature selection for allosteric binding site identification, integration of supporting amino-acid–based information to 3D structural knowledge is advantageous. This approach can enhance performance by ensuring accuracy and robustness. Therefore, we have developed an accurate and robust model called Multimodel Ensemble Feature Selection for Allosteric Site Identification (MEF-AlloSite) after collecting 9460 relevant and diverse features from the literature to characterise pockets. The model employs an accurate and robust multimodal feature selection technique for the small training set size of only 90 proteins to improve predictive performance. This state-of-the-art technique increased the performance in allosteric binding site identification by selecting promising features from 9460 features. Also, the relationship between selected features and allosteric binding sites enlightened the understanding of complex allostery for proteins by analysing selected features. MEF-AlloSite and state-of-the-art allosteric site identification methods such as PASSer2.0 and PASSerRank have been tested on three test cases 51 times with a different split of the training set. The Student’s t test and Cohen’s D value have been used to evaluate the average precision and ROC AUC score distribution. On three test cases, most of the p-values (<0.05\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$< 0.05$$\\end{document}) and the majority of Cohen’s D values (>0.5\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$> 0.5$$\\end{document}) showed that MEF-AlloSite’s 1–6% higher mean of average precision and ROC AUC than state-of-the-art allosteric site identification methods are statistically significant.

Understanding and mitigating the impact of race with adversarial autoencoders

BackgroundArtificial intelligence carries the risk of exacerbating some of our most challenging societal problems, but it also has the potential of mitigating and addressing these problems. The confounding effects of race on machine learning is an ongoing subject of research. This study aims to mitigate the impact of race on data-derived models, using an adversarial variational autoencoder (AVAE). In this study, race is a self-reported feature. Race is often excluded as an input variable, however, due to the high correlation between race and several other variables, race is still implicitly encoded in the data.MethodsWe propose building a model that (1) learns a low dimensionality latent spaces, (2) employs an adversarial training procedure that ensure its latent space does not encode race, and (3) contains necessary information for reconstructing the data. We train the autoencoder to ensure the latent space does not indirectly encode race.ResultsIn this study, AVAE successfully removes information about race from the latent space (AUC ROC = 0.5). In contrast, latent spaces constructed using other approaches still allow the reconstruction of race with high fidelity. The AVAE’s latent space does not encode race but conveys important information required to reconstruct the dataset. Furthermore, the AVAE’s latent space does not predict variables related to race (R2 = 0.003), while a model that includes race does (R2 = 0.08).ConclusionsThough we constructed a race-independent latent space, any variable could be similarly controlled. We expect AVAEs are one of many approaches that will be required to effectively manage and understand bias in ML.