SPECIAL ISSUE: VOL.-9, ISSUE-5, May 2023

Abstract

Abalone is a marine snail found in the cold coastal regions. Age is a vital characteristic that is used to determine its worth. Currently, the only viable solution to determine the age of abalone is through very detailed steps in a laboratory. This paper exploits various machine learning models for determining its age. The model was built based on a dataset obtained from the UCI Machine Learning Repository. A comprehensive analysis of various machine learning algorithms for abalone age prediction is performed which include, K-Nearest Neighbors (KNN), Naive Bayes, Decision Tree and Support Vector Machine (SVM). The three classifiers tested to evaluate their effect on its performance. Comprehensive experiments were performed using our data set.

Keywords: Abalone Age Prediction Supervised Learning Algorithms Support Vector Machine (SVM) K-Nearest Neighbors (KNN) UCI Machine Learning Repository Dataset

References

References not available

Abstract

Diabetic retinopathy, a complication of diabetes that affects the eyes, can cause damage to the blood vessels in the retina. Artificial intelligence (AI) techniques play a crucial role in computer-aided diagnosis and have proven successful in identifying various diseases. This study aims to predict diabetic retinopathy by implementing feature extraction techniques to identify relevant factors. The dataset used in this study is sourced from the UCI Machine Learning Repository. Three machine learning algorithms, namely Support Vector Machine (SVM), Multilayer Perceptron (MLP) and Naïve Bayes classifiers, are employed to analyze the dataset and determine the most effective performance and accuracy. Among these classifiers, the SVM algorithm demonstrates the highest performance with an accuracy of 96.56%.

Keywords: Diabetic Retinopathy Prediction Artificial Intelligence in Healthcare Support Vector Machine (SVM) Multilayer Perceptron (MLP) Feature Extraction Techniques

References

References not available

Abstract

Predicting heart attacks early in stroke patients through data analysis is crucial to reducing the high mortality rate associated with these conditions. However, accurately predicting heart attacks in stroke patient data poses a challenge. Early detection of stroke-related diseases is beneficial for prevention or early intervention. Machine learning and data mining play important roles in stroke prediction. This paper proposes an effective method for identifying stroke and compares the performance of three machine learning algorithms: Decision Tree, Naïve Bayes and K-Nearest Neighbors. The study analyzes the performance of these algorithms on a stroke dataset. The preliminary results demonstrate that the Decision Tree algorithm achieves the highest accuracy of 95.75%, outperforming the Naïve Bayes and K-Nearest Neighbors algorithms.

Keywords: Heart Attack Prediction in Stroke Patients Machine Learning in Healthcare Decision Tree Classifier Naïve Bayes Algorithm K-Nearest Neighbors (KNN)

References

References not available

Abstract

The examination of protein restriction locales is a significant errand in bioinformatics. Foreseeing the yeast protein restriction locales is a promising space among various exploration techniques in view of the yeast protein estimation information which have numerous records/highlights. Proteins are a significant piece of the organic entity and are engaged with pretty much every cycle in the cell. This research paper presents a comparative analysis of two supervised learning algorithms, Multilayer Perceptron (MLP) and Logistic Regression, for predicting the cellular localization sites of proteins. The algorithms were evaluated based on their accuracy, precision, and recall. The results and their implications are discussed, leading to a conclusion about the effectiveness of each algorithm in this predictive modeling task.

Keywords: Protein Cellular Localization Prediction Yeast Dataset Analysis Multilayer Perceptron (MLP) Logistic Regression Classifier Supervised Learning in Bioinformatics

References

References not available

Abstract

Pattern recognition plays a crucial role in various domains, including biology, medicine, and data analysis. Clustering algorithms, such as k-means, are commonly used for pattern recognition tasks. In this research paper, we evaluate the effectiveness of the k-means clustering algorithm on the well-known Iris dataset for identifying distinct patterns and grouping similar instances together. It highlights the evaluation of internal validation metrics, the comparison of cluster assignments with true labels, and the characteristics of the formed clusters.

Keywords: K-Means Clustering Algorithm Pattern Recognition Iris Dataset Unsupervised Learning Cluster Validation Metrics

References

References not available

Abstract

Cardiotocogram (CTG) is one of the observing apparatuses to appraise the baby wellbeing in belly. CTG for the most part yields two outcomes fetal wellbeing rate (FHR) and uterine constrictions (UC). Altogether, there are 21 ascribes in the estimation of FHR and UC on CTG. These characteristics can help obstreticians to clasify whether the embryo wellbeing is typical, thought, or neurotic. This exploration covers the discoveries and examinations of various AI models for fetal wellbeing arrangement. CTG information of 2126 pregnant ladies were gotten from the College of California Irvine AI Storehouse. Ten different AI arrangement models were prepared utilizing CTG information. Awareness, accuracy, and F1 score for each class and generally speaking precision of each model were acquired to anticipate ordinary, suspect, and neurotic fetal states. The information was inspected and utilized in a two ML models. For order, irregular woodland and it were used to cast a ballot classifier. At the point when the outcomes are analyzed, it is found that the democratic classifier model delivers the best outcomes. It accomplishes 98.62% precision, which is superior to the past technique announced.

Keywords: Fetal Health Classification Cardiotocography (CTG) Data Random Forest Classifier Voting Ensemble Model UCI Machine Learning Repository Dataset

References

References not available

Abstract

The clinical business has a great deal of data and is reliably used by researchers to cultivate new science and development to restrict the amount of passings happens due to coronary disease. Heaps of ML methods or computations are available to bring the data from informational indexes and use this got data to exactly anticipate the heart disorders. In this SPECT coronary sickness model, we used artificial intelligence computations and significant learning estimations, we have executed all computations on the dataset. The dataset used is from Kaggle which is of 267 lines and 22 trademark. The computation that are used in the model are Support Vector Machine and Multi-layer Perceptron. Hence, this paper presents a comparable report by separating the show of three man-made intelligence estimations. The primer results check that Multilayer Perceptron computation has achieved the most critical accuracy of 96.54% appeared differently in relation to Support Vector Machine This model can be valuable to the clinical specialists at their middle as decision genuinely strong organization.

Keywords: Coronary Heart Disease Prediction Artificial Intelligence in Healthcare Multilayer Perceptron (MLP) Support Vector Machine (SVM) SPECT Heart Disease Dataset

References

References not available

Abstract

In this research paper, we explore the application of machine learning algorithms, specifically Naive Bayes and K-Nearest Neighbors (KNN), for dermatology prediction. The goal is to develop a predictive model that can accurately classify dermatology conditions based on given input features. We evaluate the performance of both algorithms using metrics such as accuracy, precision, and recall. Our findings indicate promising results, demonstrating the potential of these methods in assisting dermatologists in diagnosing skin conditions effectively.

Keywords: Dermatology Disease Prediction Naïve Bayes Classification K-Nearest Neighbors (KNN) Machine Learning in Medical Diagnosis Classification Performance Metrics

References

References not available

Abstract

Hypothyroid disease is a prevalent thyroid disorder that requires accurate and early diagnosis for effective treatment. In this trial study, we investigate the effect of feature selection on the performance of a Neural Network approach for hypothyroid disease prediction. Two models are evaluated: MLP (Multi-Layer Perceptron) and MLP with SVM-RFE (Support Vector Machine - Recursive Feature Elimination). The dataset used for analysis contains relevant features as independent variables and the presence or absence of hypothyroid disease as the dependent variable.

Keywords: Hypothyroid Disease Prediction Feature Selection (SVM-RFE) Multilayer Perceptron (MLP) Neural Network Classification Machine Learning in Healthcare

References

References not available

Abstract

Malignant melanoma is a deadly form of skin cancer with a high mortality rate. To improve patient outcomes and prognosis, accurate prediction models are essential. This research paper explores the application of Support Vector Machines (SVM) and Multilayer Perceptron (MLP) algorithms to predict the prognosis of patients with malignant melanoma based on tumor measurements. The dataset comprises 205 patients who underwent tumor removal surgery. Key measurements, such as tumor thickness and ulceration status, are analyzed as potential prognostic variables. The performance of SVM and MLP algorithms in predicting patient outcomes is assessed and compared, offering insights into the effectiveness of each approach. The findings of this study have significant implications for personalized treatment strategies and patient survival rates in melanoma management.

Keywords: Malignant Melanoma Prognosis Prediction Support Vector Machine (SVM) Multilayer Perceptron (MLP) Tumor Thickness and Ulceration Analysis Machine Learning in Oncology

References

References not available

Abstract

Diabetes mellitus is a chronic metabolic disorder affecting millions of people worldwide. The increasing prevalence of diabetes poses significant challenges to healthcare systems and requires effective early detection and management strategies. This research paper explores the application of machine learning techniques for analyzing and predicting diabetes based on a comprehensive diabetic dataset. The dataset consists of various clinical and demographic features of patients, making it an ideal resource for building predictive models. Through the study, we aim to identify key factors contributing to diabetes and develop accurate models for early diagnosis. The dataset used in this study is sourced from the UCI Machine Learning Repository. Two machine learning algorithms, namely), Multilayer Perceptron (MLP) and Naïve Bayes classifiers, are employed to analyze the dataset and determine the most effective performance and accuracy. Among these classifiers, the MLP algorithm demonstrates the highest performance with an accuracy of 85.50%.

Keywords: Diabetes Mellitus Prediction Machine Learning in Healthcare Multilayer Perceptron (MLP) Naïve Bayes Classifier UCI Machine Learning Repository Dataset

References

References not available

Abstract

Bosom illness is addressed to be the most striking peril type among ladies in general and it is the second most raised ladies mishap rate among all hurtful improvement types. Conclusively expecting the diligence speed of chest disorder patients is a tremendous issue for risk scientists. AI (ML) has drawn in a great deal of thought with the presumption that it could give cautious outcomes, yet its showing systems and guess execution stay sketchy. This paper bases on the use of simulated intelligence assessments for anticipating Haberman's Bosom Malignant growth Endurance examination. Two distinct simulated intelligence approaches expressly Naïve Bayes and Logistic Regression frameworks are considered for the completion of Bosom Disease Endurance characteristic. The presentation obviously of activity of impossible to miss and typical Bosom Disease Endurance patients is assessed to the degree that various variables including arranging and testing accuracy, precision and overview. The characteristic of this deliberate outline is to see and essentially assess current appraisals concerning the usage of ML in foreseeing the 5-year constancy speed of chest destructive turn of events. Test results on Haberman's Bosom Malignant growth Endurance dataset show the force of Logistic Regression proposed system by coming to 96.73 % to the degree that exactness.

Keywords: Breast Cancer Survival Prediction Logistic Regression Naïve Bayes Classifier Haberman’s Breast Cancer Survival Dataset Machine Learning in Oncology

References

References not available

Abstract

The accurate classification of spine datasets into "Abnormal" and "Normal" classes is crucial for early diagnosis and effective treatment planning in orthopedic medicine. In this paper, we present a comparative study of two popular machine learning algorithms, K-Nearest Neighbors (KNN) and Naive Bayes, applied to a spine dataset. The objective is to determine which algorithm performs better in terms of accuracy and suitability for this specific classification task. We evaluate both methods using a dataset consisting of 310 spine samples, labeled as either "Abnormal" or "Normal." Our results demonstrate that Naive Bayes outperforms KNN, achieving an accuracy of 89%, compared to KNN's accuracy of 87%. We also discuss the implications of these findings and highlight potential areas for further research in spine dataset classification.

Keywords: Spine Abnormality Classification K-Nearest Neighbors (KNN) Naïve Bayes Algorithm Orthopedic Data Analysis Machine Learning for Medical Diagnosis

References

References not available

Abstract

Skin disorders are a critical overall clinical issue related with huge number of people. With the quick improvement of headways and the utilization of various data mining strategies of late, the progression of dermatological perceptive plan has become progressively insightful and exact. In this way, progression of computer based intelligence methodologies, which can effectively isolate dermatology ailment gathering, is basic. The motivation driving this work is to assess the presentation of computer based intelligence frameworks on skin disorders gauge utilizing Decision Tree and K-Nearest Neighbor estimations. The demonstration of the assessments is assessed through after execution assessments: accuracy, precision and review. The best outcome among two calculations for generally speaking accuracy rate was accomplished by Decision Tree model with a speed of 96.43%. This approach could improve and work with the strategy of describe the kind of skin affliction in six exceptional classes. We show that the Decision Tree performs best among others to the degree that exactness.

Keywords: Dermatology Disease Classification Decision Tree Algorithm K-Nearest Neighbor (KNN) Data Mining Techniques Machine Learning in Skin Disease Diagnosis

References

References not available

Abstract

One of the most pervasive and driving reasons for disease in ladies is bosom malignant growth. It has now turned into an incessant medical issue, and its pervasiveness has as of late expanded. The least demanding way to deal with managing bosom malignant growth discoveries is to remember them right off the bat. Accordingly, early discovery of bosom disease is basic, and with viable treatment, many lives can be saved. This examination covers the discoveries and investigations of two AI models for distinguishing bosom malignant growth. The Wisconsin Bosom Malignant growth Symptomatic dataset was utilized to foster the technique. The data was broke down and put to use in various AI models. For expectation, Random Forest and K- Nearest Neighbor were used. At the point when the outcomes are looked at, the Random Forest model is found to offer the best outcomes. Random Forest 97.54% precision, which is superior to the K-Nearest Neighbor technique.

Keywords: Breast Cancer Prediction Random Forest Algorithm K-Nearest Neighbor (KNN) Wisconsin Breast Cancer Diagnostic Dataset Machine Learning in Oncology

References

References not available

Abstract

Heart disease remains a major health concern worldwide, necessitating effective early prediction and diagnosis to reduce its impact. In this research paper, we explore the use of machine learning algorithms for heart disease prediction, leveraging data from the UCI Machine Learning Repository. The dataset contains 270 instances with 13 features and is divided into two classes: "class" with 150 instances and "Present class" with 120 instances. We employ two machine learning algorithms, Multilayer Perceptron (MLP) and Support Vector Machine (SVM), to evaluate their performance in classifying heart disease cases. Our results reveal that both MLP and SVM demonstrate promising accuracy, precision, and recall, showcasing their potential in enhancing heart disease prediction models.

Keywords: Heart Disease Prediction Multilayer Perceptron (MLP) Support Vector Machine (SVM) UCI Machine Learning Repository Dataset Machine Learning in Cardiovascular Diagnosis

References

References not available

Abstract

Cardiovascular diseases, including heart attacks, remain a significant global health concern. Early detection and accurate classification of individuals at risk are essential for effective prevention and intervention. In this research paper, we evaluate the performance of two popular machine learning algorithms, K-Nearest Neighbors (K-NN) and Naïve Bayes, for the classification of heart attack cases. Using a dataset comprising 1319 instances and 9 features, we compare the accuracy, precision, and recall of these algorithms.

Keywords: Heart Attack Classification K-Nearest Neighbors (K-NN) Naïve Bayes Classifier Cardiovascular Disease Prediction Machine Learning in Healthcare

References

References not available

Abstract

This research paper delves into the intricate task of heart disease diagnosis, utilizing a multivariate dataset encompassing 14 distinct attributes. The dataset, commonly referred to as the Cleveland database, has been widely adopted in machine learning research. With 606 instances and attributes spanning age, sex, medical parameters, and electrocardiographic results, the primary objectives of this study are two-fold: Firstly, to develop predictive models that accurately identify the presence of heart disease based on patient attributes, and secondly, to unearth valuable insights from the dataset that contribute to a deeper understanding of this critical health concern. Two classification algorithms, Naive Bayes and Logistic Regression, are employed and their performance in terms of accuracy, precision, and recall are compared.

Keywords: Heart Disease Diagnosis Naïve Bayes Classifier Logistic Regression Cleveland Heart Disease Dataset Multivariate Data Analysis

References

References not available

Abstract

Liver disease is a major health concern worldwide, and early diagnosis is crucial for effective treatment and management. In this research, we employ ensemble classification techniques to predict the presence or absence of liver disease using a dataset comprising 441 male and 142 female patient records. We compare the performance of two popular ensemble algorithms, AdaBoost and Gradient Boosting, in terms of accuracy, precision, and recall. Our results demonstrate that both AdaBoost and Gradient Boosting exhibit high accuracy, precision, and recall rates, making them promising tools for liver disease prediction. This research contributes to the growing body of literature on ensemble classification methods for medical diagnosis and highlights the potential of these techniques in improving healthcare outcomes.

Keywords: Liver Disease Prediction Ensemble Learning AdaBoost Algorithm Gradient Boosting Algorithm Machine Learning in Medical Diagnosis

References

References not available

Abstract

This research paper addresses the challenging task of multiclass classification within the realm of dermatology, employing the versatile Logistic Regression algorithm. The dataset under investigation comprises six distinct classes representing various dermatological disorders. The primary objectives of this study are twofold: to evaluate the performance of Logistic Regression when implemented with the One vs One and One vs Rest strategies, and to assess their effectiveness in classifying dermatological conditions accurately. Results indicate that both strategies exhibit remarkable classification accuracy, precision, and recall rates, underscoring their potential in dermatological diagnosis.

Keywords: Dermatology Multiclass Classification Logistic Regression Algorithm One-vs-One (OvO) Strategy One-vs-Rest (OvR) Strategy Machine Learning in Skin Disease Diagnosis

References

References not available

Abstract

In the field of medical data analysis, the accurate classification of primary tumor surgeries is of paramount importance for diagnosing and treating patients effectively. In this study, we explore the application of Logistic Regression with two different multilabel strategies, namely one-vs-one and one-against-one, to predict primary tumor surgery outcomes using the Primary Tumor Surgery dataset. This dataset consists of 339 data samples with 18 features and 21 distinct classes. Our experiments reveal promising results, with the one-vs-one approach achieving an accuracy of 93.67%, precision of 93.7%, and recall of 93.7%, while the one-against-one approach attained an accuracy of 92.45%, precision of 92.4%, and recall of 92.5%. This research not only highlights the effectiveness of Logistic Regression for multilabel prediction in the medical domain but also emphasizes the significance of choosing an appropriate multilabel strategy for optimizing classification performance. We discuss the implications of our findings and potential applications in improving patient care.

Keywords: Primary Tumor Surgery Classification Multilabel Classification Logistic Regression One-vs-One Strategy Medical Data Analysis

References

References not available

Abstract

Thyroid disease is a prevalent endocrine disorder affecting millions of people worldwide. Timely diagnosis and accurate prediction of thyroid disease are crucial for effective patient care. In this research paper, we investigate the performance of two popular machine learning algorithms, Multilayer Perceptron (MLP) and Support Vector Machine (SVM), in predicting thyroid disease based on a comprehensive dataset containing 30 attributes and 3772 instances with two class labels: Negative and Sick. Our results indicate that MLP achieved superior predictive accuracy, precision, and recall compared to SVM, with an accuracy rate of 96.76%, precision of 96.7%, and recall of 96.8%. These findings suggest that MLP may be a valuable tool for improving thyroid disease diagnosis and patient outcomes. This paper discusses the implications of these results for clinical practice and future research directions.

Keywords: Thyroid Disease Prediction Multilayer Perceptron (MLP) Support Vector Machine (SVM) Endocrine Disorder Classification Machine Learning in Healthcare

References

References not available

Abstract

This research paper presents a comprehensive evaluation of the decision tree algorithm using two attribute selection measures, namely Gini (information gain) and Entropy, for the classification of Breast Cancer Wisconsin (Diagnostic) Data. The primary objectives were to assess the number of selected features for the root node and the resulting learning accuracy. The study found that while both attribute selection measures yield promising results, Entropy outperforms Gini in terms of accuracy and precision. This research sheds light on the importance of feature selection in machine learning models for medical diagnosis.

Keywords: Breast Cancer Wisconsin (Diagnostic) Dataset Decision Tree Algorithm Gini Index (Information Gain) Entropy-Based Attribute Selection Feature Selection in Medical Diagnosis

References

References not available