A Comprehensive Investigation of the Performances of Different Machine Learning Classifiers with SMOTE-ENN Oversampling Technique and Hyperparameter Optimization for Imbalanced Heart Failure Dataset

Scientific Programming, Journal Year: 2022, Volume and Issue: 2022, P. 1 - 17

Published: March 9, 2022

Heart failure is a chronic cardiac condition characterized by reduced supply of blood to the body due impaired contractile properties muscles heart. Like any other disorder, heart serious ailment limiting activities and curtailing lifespan patient, most often resulting in death sooner or later. Detection survival patients with path effective intervention good prognosis terms both treatment quality life patient. Machine learning techniques can be critical this regard since they used predict advance, allowing receive appropriate treatment. Hence, six supervised machine algorithms have been studied applied analyze dataset 299 individuals from UCI Learning Repository their survivability failure. Three distinct approaches followed using Decision Tree Classifier, Logistic Regression, Gaussian Naïve Bayes, Random Forest K-Nearest Neighbors, Support Vector algorithms. Data scaling has performed as preprocessing step utilizing standard min–max method. However, grid search cross-validation random employed optimize hyperparameters. Additionally, synthetic minority oversampling technique edited nearest neighbor (SMOTE-ENN) data resampling are utilized, performances all compared extensively. The experimental results clearly indicate that Classifier (RFC) surpasses test accuracy 90% when combination SMOTE-ENN technique. Therefore, comprehensive investigation portrays vivid visualization applicability compatibility different such an imbalanced presents role algorithm hyperparameter optimization for enhancing

Language: Английский

---

A comparative assessment of artificial intelligence models used for early prediction and evaluation of chronic kidney disease

Decision Analytics Journal, Journal Year: 2023, Volume and Issue: 6, P. 100169 - 100169

Published: Jan. 26, 2023

Chronic Kidney Disease (CKD) is one of the most prevalent and fatal diseases influencing people on a larger that remains dormant until irreversible damage has been done to an individual's kidney. Progression CKD related variety great complications, including increased incidence various disorders, anemia, hyperlipidemia, nerve damage, pregnancy complication, even complete kidney failure. Millions die from this disease every year. Diagnosing cumbersome task as no major symptoms can be used benchmark detect disease. In cases when diagnosis persists, some results may interpreted incorrectly. This study proposes using deep neural network-based Multi-Layer Perceptron Classifier diagnose in patients. The model was trained data 400 considered signs, age, blood sugar, red cell count, etc. Experiments reveal proposed achieves perfect testing accuracy classification tasks. Our goal facilitate introducing Deep Learning approaches learning dataset attribute reports accurately detecting CKD. paper's primary contribution Neural Network for chronic 100% accuracy, outperforming standard machine models such support vector machines naive Bayes classifiers. paper provides detailed explanation multi-layer perceptron classifier, which uses network provided by PyTorch library its basis. better alternative adaption techniques classifying Because they handle non-linearity data, compute complex heaps fetched datasets, adapt learn their own about key information layers neurons present structure.

Language: Английский

---

A Robust Chronic Kidney Disease Classifier Using Machine Learning

Electronics, Journal Year: 2023, Volume and Issue: 12(1), P. 212 - 212

Published: Jan. 1, 2023

Clinical support systems are affected by the issue of high variance in terms chronic disorder prognosis. This uncertainty is one principal causes for demise large populations around world suffering from some fatal diseases such as kidney disease (CKD). Due to this reason, diagnosis great concern healthcare systems. In a case, machine learning can be used an effective tool reduce randomness clinical decision making. Conventional methods detection not always accurate because their degree dependency on several sets biological attributes. Machine process training using vast collection historical data purpose intelligent classification. work aims at developing machine-learning model that use publicly available forecast occurrence disease. A set preprocessing steps were performed dataset order construct generic model. includes appropriate imputation missing points, along with balancing SMOTE algorithm and scaling features. statistical technique, namely, chi-squared test, extraction least-required adequate highly correlated features output. For training, stack supervised-learning techniques development robust Out all applied techniques, vector (SVM) random forest (RF) achieved lowest false-negative rates test accuracy, equal 99.33% 98.67%, respectively. However, SVM better results than RF did when validated 10-fold cross-validation.

Language: Английский

---

Machine Learning Hybrid Model for the Prediction of Chronic Kidney Disease

Computational Intelligence and Neuroscience, Journal Year: 2023, Volume and Issue: 2023, P. 1 - 14

Published: March 14, 2023

To diagnose an illness in healthcare, doctors typically conduct physical exams and review the patient's medical history, followed by diagnostic tests procedures to determine underlying cause of symptoms. Chronic kidney disease (CKD) is currently leading death, with a rapidly increasing number patients, resulting 1.7 million deaths annually. While various methods are available, this study utilizes machine learning due its high accuracy. In study, we have used hybrid technique build our proposed model. model, Pearson correlation for feature selection. first step, best models were selected on basis critical literature analysis. second combination these Gaussian Naïve Bayes, gradient boosting, decision tree classifier as base classifier, random forest meta-classifier The objective evaluate classification techniques identify best-used terms This provides solution overfitting achieves highest It also highlights some challenges that affect result better performance. critically existing available techniques. We accuracy, comprehensive analytical evaluation related work presented tabular system. implementation, top four built model using UCI chronic dataset prediction. Gradient boosting around 99% 98%, 96% performs getting 100% accuracy same dataset. Some main algorithms predict occurrence CKD tree, K-nearest neighbor, forest, support vector machine, LDA, GB, neural network. apply GB (gradient boosting), along set features compare score.

Language: Английский

---

Enhancing the Early Detection of Chronic Kidney Disease: A Robust Machine Learning Model

Big Data and Cognitive Computing, Journal Year: 2023, Volume and Issue: 7(3), P. 144 - 144

Published: Aug. 16, 2023

Clinical decision-making in chronic disorder prognosis is often hampered by high variance, leading to uncertainty and negative outcomes, especially cases such as kidney disease (CKD). Machine learning (ML) techniques have emerged valuable tools for reducing randomness enhancing clinical decision-making. However, conventional methods CKD detection lack accuracy due their reliance on limited sets of biological attributes. This research proposes a novel ML model predicting CKD, incorporating various preprocessing steps, feature selection, hyperparameter optimization technique, algorithms. To address challenges medical datasets, we employ iterative imputation missing values sequential approach data scaling, combining robust z-standardization, min-max scaling. Feature selection performed using the Boruta algorithm, developed The proposed was validated UCI dataset, achieving outstanding performance with 100% accuracy. Our approach, innovative k-nearest neighbors along grid-search cross-validation (CV), demonstrates its effectiveness early CKD. highlights potential improving support systems impact prognosis.

Language: Английский

---

Toward Comprehensive Chronic Kidney Disease Prediction Based on Ensemble Deep Learning Models

Applied Sciences, Journal Year: 2023, Volume and Issue: 13(6), P. 3937 - 3937

Published: March 20, 2023

Chronic kidney disease (CKD) refers to the gradual decline of function over months or years. Early detection CKD is crucial and significantly affects a patient’s decreasing health progression through several methods, including pharmacological intervention in mild cases hemodialysis transportation severe cases. In recent past, machine learning (ML) deep (DL) models have become important medical diagnosis domain due their high prediction accuracy. The performance developed model mainly depends on choosing appropriate features suitable algorithms. Accordingly, paper aims introduce novel ensemble DL approach detect CKD; multiple methods feature selection were used select optimal selected features. Moreover, we study effect chosen from side. proposed integrates pretrained with support vector (SVM) as metalearner model. Extensive experiments conducted by using 400 patients UCI repository. results demonstrate efficiency compared other models. mutual_info_classi obtained highest performance.

Language: Английский

---

On the diagnosis of chronic kidney disease using a machine learning-based interface with explainable artificial intelligence

Intelligent Systems with Applications, Journal Year: 2024, Volume and Issue: 22, P. 200397 - 200397

Published: June 1, 2024

Chronic Kidney Disease (CKD) is increasingly recognised as a major health concern due to its rising prevalence. The average survival period without functioning kidneys typically limited approximately 18 days, creating significant need for kidney transplants and dialysis. Early detection of CKD crucial, machine learning methods have proven effective in diagnosing the condition, despite their often opaque decision-making processes. This study utilised explainable predict CKD, thereby overcoming 'black box' nature traditional predictions. Of six algorithms evaluated, extreme gradient boost (XGB) demonstrated highest accuracy. For interpretability, employed Shapley Additive Explanations (SHAP) Partial Dependency Plots (PDP), which elucidate rationale behind predictions support process. Moreover, first time, graphical user interface with explanations was developed diagnose likelihood CKD. Given critical high stakes use can aid healthcare professionals making accurate diagnoses identifying root causes.

Language: Английский

---

Detection of Parkinson's Disease by Employing Boosting Algorithms

Published: Aug. 16, 2021

Parkinson's Disease is caused by a decline in the production of dopamine due to degeneration brain cells. Dopamine responsible for communication between parts associated with control and fluency body movements. Hence, disease manifests spectrum movement disorders as well non-motor features. It now revealed that symptoms may show many years prior onset motor symptoms. Therefore, early accurate diagnosis crucial stop or slow down progression its tracks. In this context, ensemble machine learning (ML) algorithms like boosting can play significant role detecting at an stage. paper, four are studied implemented UCI dataset. After rigorous simulation, ML models exhibited satisfactory results terms different performance parameters accuracy, precision, recall, F1-Score., AUC, Youden, specificity error rate. However, performances model improved tuning hyperparameters GridSearchCV. detailed comparative analysis portrayed where Light GBM displayed highest accuracy 93.39% after hyperparameter tuning. XGBoost Gradient Boosting algorithm also depicted accuracies more than 90% but AdaBoost demonstrated maximum 87.22%

Language: Английский

---

Early risk prediction of cervical cancer: A machine learning approach

2022 19th International Conference on Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON), Journal Year: 2022, Volume and Issue: unknown, P. 1 - 4

Published: May 24, 2022

Cervical cancer is a vital public health issue that affects women worldwide. As it fatal disease, early risk prediction of cervical can play an important role in prevention by raising awareness this disease. Early using Machine Learning (ML) model be beneficial solution for both healthcare professionals and people at risk. In study, eleven supervised ML algorithms are utilized to forecast jeopardies disease dataset from UCI repository. The models rummaged prophesy the threats, performance parameters like accuracy, precision, F1-score, re-call, ROC-AUC estimated. Finally, reasonable analysis performed, revealing study achieved 93.33% accuracy with Multi-Layer Perceptron (MLP) algorithm default hyperparameters. However, employing hyperparameter tuning method Grid Search Cross Validation (GSCV), K-Nearest Neighbors (KNN), Decision Tree Classifier (DTC), Support Vector (SVM), Random Forest (RFC), all portrayed 93.33%.

Language: Английский

---

A modified weighted mean of vectors optimizer for Chronic Kidney disease classification

Computers in Biology and Medicine, Journal Year: 2023, Volume and Issue: 155, P. 106691 - 106691

Published: Feb. 16, 2023

Language: Английский

---