Revolutionizing core muscle analysis in female sexual dysfunction based on machine learning

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Фев. 27, 2024

Abstract The purpose of this study is to investigate the role core muscles in female sexual dysfunction (FSD) and develop comprehensive rehabilitation programs address issue. We aim answer following research questions: what are roles FSD, how can machine deep learning models accurately predict changes during FSD? FSD a common condition that affects women all ages, characterized by symptoms such as decreased libido, difficulty achieving orgasm, pain intercourse. conducted analysis using learning. evaluated performance multiple models, including multi-layer perceptron (MLP), long short-term memory (LSTM), convolutional neural network (CNN), recurrent (RNN), ElasticNetCV, random forest regressor, SVR, Bagging regressor. were based on mean squared error (MSE), absolute (MAE), R-squared (R 2 ) score. Our results show CNN regressor most accurate for predicting FSD. achieved lowest MSE (0.002) highest R score (0.988), while also performed well with an 0.0021 0.9905. demonstrates neglected play significant highlighting need these muscles. By developing programs, we improve quality life help them achieve optimal health.

Язык: Английский

---

Employing machine learning for enhanced abdominal fat prediction in cavitation post-treatment

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Май 14, 2024

This study investigates the application of cavitation in non-invasive abdominal fat reduction and body contouring, a topic considerable interest medical aesthetic fields. We explore potential to alter composition delve into optimization prediction models using advanced hyperparameter techniques, Hyperopt Optuna. Our objective is enhance predictive accuracy dynamics post-cavitation treatment. Employing robust dataset with measurements treatment parameters, we evaluate efficacy our approach through regression analysis. The performance Optuna assessed metrics such as mean squared error, absolute R-squared score. results reveal that both exhibit strong capabilities, scores reaching 94.12% 94.11% for post-treatment visceral fat, 71.15% 70.48% subcutaneous predictions, respectively. Additionally, investigate feature selection techniques pinpoint critical predictors within models. Techniques including F-value selection, mutual information, recursive elimination logistic random forests, variance thresholding, importance evaluation are utilized. analysis identifies key features BMI, waist circumference, pretreatment levels significant outcomes. findings underscore effectiveness refining offer valuable insights advancement methods. research holds important implications scientific community clinical practitioners, paving way improved strategies realm contouring.

Язык: Английский

---

Utilizing machine learning to analyze trunk movement patterns in women with postpartum low back pain

Scientific Reports, Год журнала: 2024, Номер 14(1)

Опубликована: Авг. 12, 2024

Abstract This paper presents an analysis of trunk movement in women with postnatal low back pain using machine learning techniques. The study aims to identify the most important features related and develop accurate models for predicting pain. Machine approaches showed promise analyzing biomechanical factors (LBP). applied regression classification algorithms proposed dataset from 100 postpartum women, 50 LBP without. Optimized optuna Regressor achieved best performance a mean squared error (MSE) 0.000273, absolute (MAE) 0.0039, R2 score 0.9968. In classification, Basic CNN Random Forest Classifier both attained near-perfect accuracy 1.0, area under receiver operating characteristic curve (AUC) precision recall F1-score outperforming other models. Key predictive included (correlation -0.732 flexion range motion), motion measures (flexion extension correlation 0.662), average movements 0.957 flexion). Feature selection consistently identified pain, flexion, extension, lateral as influential across methods. While limited this initial constrained by generalizability, offered quantitative insight. Models accurately regressed (MSE < 0.01, > 0.95) classified (accuracy 0.94) biomechanics distinguishing LBP. Incorporating additional demographic, clinical, patient-reported may enhance individualized risk prediction treatment personalization. preliminary application advanced analytics supported learning's potential utility determination outcome improvement. provides valuable insights into use techniques can potentially inform development more effective treatments. Trial registration : trial was designed observational cross-section study. approved Ethical Committee Deraya University, Faculty Pharmacy, (No: 10/2023). According ethical standards Declaration Helsinki. complies principles human research. Each patient signed written consent form after being given thorough description trial. conducted at outpatient clinic February 2023 till June 30, 2023.

Язык: Английский

---

Enhancing Heart Attack Prediction with Machine Learning: A Study at Jordan University Hospital

Applied Computational Intelligence and Soft Computing, Год журнала: 2024, Номер 2024(1)

Опубликована: Янв. 1, 2024

Efforts have been made to address the adverse impact of heart disease on society by improving its treatment and diagnosis. This study uses Jordan University Hospital (JUH) Heart Dataset develop evaluate machine‐learning models for predicting disease. The primary objective this is enhance prediction accuracy utilizing a comprehensive approach that includes data preprocessing, feature selection, model development. Various artificial intelligence techniques, namely, random forest, SVM, decision tree, naive Bayes, K‐nearest neighbours (KNN) were explored with particle swarm optimization (PSO) selection. These results substantial implications early detection, diagnosis, tailored treatment, potentially aiding medical professionals in making well‐informed decisions patient outcomes. PSO used select most compelling features out 58 features. Experiments dataset comprising 486 patients at JUH yielded commendable classification 94.3% using our proposed system, aligning state‐of‐the‐art performance. Notably, research utilized distinct provided corresponding author, while alternative algorithms achieved accuracies ranging from 85% 90%. emphasize superior system compared other considered, particularly highlighting SVM classifier as accurate, contributing significantly diagnosis regions like Jordan, where cardiovascular diseases are leading cause mortality.

Язык: Английский

---

Machine Learning Methods in Physical Therapy: A Scoping Review of applications in clinical context.

Musculoskeletal Science and Practice, Год журнала: 2024, Номер 74, С. 103184 - 103184

Опубликована: Сен. 13, 2024

Язык: Английский

---

Optimal reconfiguration of distribution systems considering reliability: Introducing long-term memory component AEO algorithm

Expert Systems with Applications, Год журнала: 2024, Номер 249, С. 123467 - 123467

Опубликована: Фев. 15, 2024

This article introduces a modified version of the Artificial Ecosystem Optimization (AEO) algorithm, called Long-term Memory Component AEO (LMAEO), for optimizing reconfiguration radial distribution networks. The LMAEO algorithm incorporates long-term memory component, enabling individuals in population to make decisions based on past experiences. integration allows explore wider range potential solutions during optimization process, potentially leading improved performance and better exploration solution space. To verify effectiveness superiority technique, it is compared with conventional other well-known algorithms using seven benchmark functions. proposed successfully addresses systems considering reliability 12-bus, 33-bus 69-bus IEEE test systems. Leveraging strengths achieves efficient this problem. assess LMAEO, comparison made original algorithm. results demonstrate that technique surpasses optimizer terms optimal jointly reliability, system losses voltage deviations.

Язык: Английский

---

Machine Learning–Based Hyperglycemia Prediction: Enhancing Risk Assessment in a Cohort of Undiagnosed Individuals

JMIRx Med, Год журнала: 2024, Номер 5, С. e56993 - e56993

Опубликована: Апрель 24, 2024

Noncommunicable diseases continue to pose a substantial health challenge globally, with hyperglycemia serving as prominent indicator of diabetes.

Язык: Английский

---

Optimization of diabetes prediction methods based on combinatorial balancing algorithm

Nutrition and Diabetes, Год журнала: 2024, Номер 14(1)

Опубликована: Авг. 14, 2024

Diabetes, as a significant disease affecting public health, requires early detection for effective management and intervention. However, imbalanced datasets pose challenge to accurate diabetes prediction. This imbalance often results in models performing poorly predicting minority classes, overall diagnostic performance. To address this issue, study employs combination of Synthetic Minority Over-sampling Technique (SMOTE) Random Under-Sampling (RUS) data balancing uses Optuna hyperparameter optimization machine learning models. approach aims fill the gap current research concerning model optimization, thereby improving prediction accuracy computational efficiency. First, SMOTE RUS methods process dataset, distribution. Then, is utilized optimize hyperparameters LightGBM enhance its During experiment, effectiveness proposed evaluated by comparing training dataset before after balancing. The experimental show that enhanced LightGBM-Optuna improves from 97.07% 97.11%, precision 97.17% 98.99%. time required single search only 2.5 seconds. These demonstrate superiority method handling optimizing indicates combining algorithms with can effectively models, especially dealing

Язык: Английский

---

Development and validation of a new nomogram for self-reported OA based on machine learning: a cross-sectional study

Scientific Reports, Год журнала: 2025, Номер 15(1)

Опубликована: Янв. 4, 2025

Developing a new diagnostic prediction model for osteoarthritis (OA) to assess the likelihood of individuals developing OA is crucial timely identification potential populations OA. This allows further diagnosis and intervention, which significant improving patient prognosis. Based on NHANES periods 2011–2012, 2013–2014, 2015–2016, study involved 11,366 participants, whom 1,434 reported LASSO regression, XGBoost algorithm, RF algorithm were used identify indicators, nomogram was developed. The evaluated by measuring AUC, calibration curve, DCA curve training validation sets. In this study, we identified 5 predictors from 19 variables, including age, gender, hypertension, BMI caffeine intake, developed an nomogram. both cohorts, exhibited good predictive performance (with AUCs 0.804 0.814, respectively), consistency stability in high net benefit DCA. based variables demonstrates accuracy predicting OA, indicating that it convenient tool clinicians

Язык: Английский

---

Novel approach for noninvasive pelvic floor muscle strength measurement using extracorporeal surface perineal pressure measurement and machine learning modeling

Digital Health, Год журнала: 2025, Номер 11

Опубликована: Янв. 1, 2025

Objective Accurate measurement of pelvic floor muscle (PFM) strength is crucial for the management disorders. However, current methods are invasive, uncomfortable, and lack standardization. This study aimed to introduce a novel noninvasive approach precise PFM quantification by leveraging extracorporeal surface perineal pressure (ESPP) measurements machine learning algorithms. Methods Twenty-one healthy women participated in this study. ESPP were obtained using 10 × array sensor during maximal voluntary contractions seated position. Simultaneously, transabdominal ultrasound was used measure bladder base displacement (mm) as reference contraction strength. Seven variables calculated based on data intra- inter-rater reliabilities assessed. Machine algorithms predicted from variables. Results The demonstrated good excellent intra-rater (ICC = 0.881) 0.967) reliability. Significant correlations observed between middle ( r .619, P < .001) front −.379, =.002) vectors. top-performing models predicting support vector [root mean square error (RMSE) 0.139, R2 0.542], random forest (RMSE 0.123, 0.367), AdaBoost 0.320) training set, 0.173, 0.537), 0.177, 0.512), 0.178, 0.508) test set. In displacement, Bland–Altman analysis revealed these had minimal systematic bias, with differences ranging −0.007 0.066, clinically acceptable limits agreement. Conclusion demonstrates potential reliable valid assessing quantifying directionality contractions, overcoming limitations traditional techniques.

Язык: Английский

---