MedAi: A Smartwatch-Based Application Framework for the Prediction of Common Diseases Using Machine Learning DOI Creative Commons
Shinthi Tasnim Himi, Natasha Tanzila Monalisa, Md Whaiduzzaman

et al.

IEEE Access, Journal Year: 2023, Volume and Issue: 11, P. 12342 - 12359

Published: Jan. 1, 2023

Health information technology is one of today's fastest-growing and most powerful technologies. This used predominantly for predicting illness obtaining medications quickly because visiting a doctor performing pathological tests can be time-consuming expensive. has prompted many researchers to contribute by developing new disease prediction systems or improving existing ones. paper presents smartwatch-based system named 'MedAi' multiple diseases such as ischemic heart disease, hypertension, respiratory hyperthyroidism, hypothyroidism, stroke, myocardial infarction, kidney failure, gallstones, diabetes, dyslipidemia using machine learning algorithms. It comprises three core modules: prototype smartwatch 'Sense O'Clock' equipped with eleven sensors collect bodily statistics, model analyze the data make prediction, mobile application display result. A dataset consisting patient statistics was obtained from local hospital according ethical guidelines, prior consent both patients doctors. We employ several algorithms, including Support Vector Machine (SVM), Regression (SVR), K-Nearest Neighbor (KNN), Extreme Gradient Boosting (XGBoost), Long Short Term Memory (LSTM), Random Forest (RF) investigate best algorithm. Experimentation our shows that RF algorithm outperforms other algorithms SVM, KNN, XGBoost, etc., in aforementioned an accuracy 99.4%. The provides full-time assistance user reporting his her body condition suggesting requisite remedies. notable addition early predict vulnerabilities before they reach irrecoverable stage. Finally, we compare method related methods.

Language: Английский

A Novel Deep Learning-Based Intrusion Detection System for IoT Networks DOI Creative Commons
Albara Awajan

Computers, Journal Year: 2023, Volume and Issue: 12(2), P. 34 - 34

Published: Feb. 5, 2023

The impressive growth rate of the Internet Things (IoT) has drawn attention cybercriminals more than ever. growing number cyber-attacks on IoT devices and intermediate communication media backs claim. Attacks IoT, if they remain undetected for an extended period, cause severe service interruption resulting in financial loss. It also imposes threat identity protection. Detecting intrusion real-time is essential to make IoT-enabled services reliable, secure, profitable. This paper presents a novel Deep Learning (DL)-based detection system devices. intelligent uses four-layer deep Fully Connected (FC) network architecture detect malicious traffic that may initiate attacks connected proposed been developed as protocol-independent reduce deployment complexities. demonstrates reliable performance simulated real intrusions during experimental analysis. detects Blackhole, Distributed Denial Service, Opportunistic Sinkhole, Workhole with average accuracy 93.74%. system’s precision, recall, F1-score are 93.71%, 93.82%, 93.47%, respectively, average. innovative learning-based IDS maintains 93.21% which satisfactory improving security networks.

Language: Английский

Citations

123
Robust clinical applicable CNN and U-Net based algorithm for MRI classification and segmentation for brain tumor DOI Creative Commons
Atika Akter,

Nazeela Nosheen,

Sabbir Ahmed

et al.

Expert Systems with Applications, Journal Year: 2023, Volume and Issue: 238, P. 122347 - 122347

Published: Oct. 28, 2023

Early diagnosis of brain tumors is critical for enhancing patient prognosis and treatment options, while accurate classification segmentation are vital developing personalized strategies. Despite the widespread use Magnetic Resonance Imaging (MRI) examination advances in AI-based detection methods, building an efficient model detecting categorizing from MRI images remains a challenge. To address this problem, we proposed deep Convolutional Neural Network (CNN)-based architecture automatic image into four classes U-Net-based model. Using six benchmarked datasets, tested trained model, enabling side-by-side comparison impact on tumor images. We also evaluated two methods based accuracy, recall, precision, AUC. Our developed novel learning-based outperforms existing pre-trained models across all datasets. The results demonstrate that our achieved highest accuracy 98.7% merged dataset 98.8% with approach, reaching 97.7% among individual Thus, framework could be applicable clinics identification utilizing scan input

Language: Английский

Citations

93
Intelligent Hybrid Deep Learning Model for Breast Cancer Detection DOI Open Access
Xiaomei Wang, Ijaz Ahmad, Danish Javeed

et al.

Electronics, Journal Year: 2022, Volume and Issue: 11(17), P. 2767 - 2767

Published: Sept. 2, 2022

Breast cancer (BC) is a type of tumor that develops in the breast cells and one most common cancers women. Women are also at risk from BC, second life-threatening disease after lung cancer. The early diagnosis classification BC very important. Furthermore, manual detection time-consuming, laborious work, and, possibility pathologist errors, incorrect classification. To address above highlighted issues, this paper presents hybrid deep learning (CNN-GRU) model for automatic BC-IDC (+,−) using whole slide images (WSIs) well-known PCam Kaggle dataset. In research, proposed used different layers architectures CNNs GRU to detect IDC validation tests quantitative results were carried out each performance measure (accuracy (Acc), precision (Prec), sensitivity (Sens), specificity (Spec), AUC F1-Score. shows best measures 86.21%, 85.50%, 85.60%, 84.71%, F1-score 88%, while 0.89 which overcomes pathologist’s error miss problem. Additionally, efficiency was tested compared with CNN-BiLSTM, CNN-LSTM, current machine (ML/DL) models, indicated more robust than recent ML/DL approaches.

Language: Английский

Citations

81
En-MinWhale: An Ensemble Approach Based on MRMR and Whale Optimization for Cancer Diagnosis DOI Creative Commons
Amrutanshu Panigrahi, Abhilash Pati, Bibhuprasad Sahu

et al.

IEEE Access, Journal Year: 2023, Volume and Issue: 11, P. 113526 - 113542

Published: Jan. 1, 2023

According to the WHO, Cancer is a prominent cause of mortality worldwide, accounting for ~ 10 million fatalities at end 2020. The most common types cancers include Lung, Breast, CNS, Leukemia, Colon, and Cervical Cancer. Early detection cancer can decrease death toll. study, if identified its early stage, rate be reduced ~85%. In order reduce toll, machine learning (ML) emerges as significant solution. When it comes research with ML, biopsy microarray data come into front. less useful excludes patient's genetic information. However, due information, solution detecting disease. Dealing also has some consequences, high dimensionality one them. This article reports an ML-based ensemble model tackle issues provide effective detection. reported uses Minimum Redundancy Maximum Relevance (MRMR) feature selection algorithm. Whale Optimization Algorithm (WOA) implemented featured dataset select optimistic number features without affecting relevance. Then, four classification models, including Support Vector Machine, Decision Tree, Multi-Layer Perceptron, Random Forest, are applied base learners make initial predictions. Finally, voting technique prediction develop prediction. proposed En-MinWhale evaluated over six different datasets, Ovarian, Colon performance using 11 various evaluative parameters, accuracy, precision, specificity, sensitivity, F-β score, etc. shows 94.09%, 95.83%, 94.86%, 95.00%, 94.85%, 96.77% accuracy respectively, that outperforms other considered hybrid models help out physicians in diagnosis.

Language: Английский

Citations

54
Lung cancer detection from CT scans using modified DenseNet with feature selection methods and ML classifiers DOI
Madhusudan G. Lanjewar, Kamini G. Panchbhai, Charanarur Panem

et al.

Expert Systems with Applications, Journal Year: 2023, Volume and Issue: 224, P. 119961 - 119961

Published: March 25, 2023

Language: Английский

Citations

53
Synergizing the enhanced RIME with fuzzy K-nearest neighbor for diagnose of pulmonary hypertension DOI

Xiao-Ming Yu,

Wenxiang Qin,

Xiao Lin

et al.

Computers in Biology and Medicine, Journal Year: 2023, Volume and Issue: 165, P. 107408 - 107408

Published: Aug. 29, 2023

Language: Английский

Citations

47
SafetyMed: A Novel IoMT Intrusion Detection System Using CNN-LSTM Hybridization DOI Open Access
Nuruzzaman Faruqui, Mohammad Abu Yousuf, Md Whaiduzzaman

et al.

Electronics, Journal Year: 2023, Volume and Issue: 12(17), P. 3541 - 3541

Published: Aug. 22, 2023

The Internet of Medical Things (IoMT) has become an attractive playground to cybercriminals because its market worth and rapid growth. These devices have limited computational capabilities, which ensure minimum power absorption. Moreover, the manufacturers use simplified architecture offer a competitive price in market. As result, IoMTs cannot employ advanced security algorithms defend against cyber-attacks. IoMT easy prey for due access valuable data rapidly expanding market, as well being comparatively easier exploit.As intrusion rate is experiencing surge. This paper proposes novel Intrusion Detection System (IDS), namely SafetyMed, combining Convolutional Neural Networks (CNN) Long Short-Term Memory (LSTM) networks from sequential grid data. SafetyMed first IDS that protects malicious image network traffic. innovative ensures optimized detection by trade-off between False Positive Rate (FPR) (DR). It detects intrusions with average accuracy 97.63% precision recall, F1-score 98.47%, 97%, 97.73%, respectively. In summary, potential revolutionize many vulnerable sectors (e.g., medical) ensuring maximum protection intrusion.

Language: Английский

Citations

44
A Review of Emerging Technologies for IoT-Based Smart Cities DOI Creative Commons
Md Whaiduzzaman, Alistair Barros,

Moumita Chanda

et al.

Sensors, Journal Year: 2022, Volume and Issue: 22(23), P. 9271 - 9271

Published: Nov. 28, 2022

Smart cities can be complemented by fusing various components and incorporating recent emerging technologies. IoT communications are crucial to smart city operations, which designed support the concept of a “Smart City” utilising most cutting-edge communication technologies enhance administration resident services. have been outfitted with numerous IoT-based gadgets; Internet Things is modular method integrate sensors all ICT This paper provides an overview cities’ concepts, characteristics, applications. We thoroughly investigate applications, challenges, possibilities solutions in technological trends perspectives, such as machine learning blockchain. discuss cloud fog ecosystems capacity devices, architectures, approaches. In addition we security privacy aspects, including blockchain towards more trustworthy resilient cities. also highlight applications provide conceptual model mega-events framework. Finally, outline impact technologies’ implications on for futuristic

Language: Английский

Citations

59
HRDEL: High ranking deep ensemble learning-based lung cancer diagnosis model DOI

Kanchan Pradhan,

Priyanka Chawla, Rajeev Tiwari

et al.

Expert Systems with Applications, Journal Year: 2022, Volume and Issue: 213, P. 118956 - 118956

Published: Oct. 5, 2022

Language: Английский

Citations

45
Recent advancements in deep learning based lung cancer detection: A systematic review DOI
Shubham Dodia, B. Annappa, P A Mahesh

et al.

Engineering Applications of Artificial Intelligence, Journal Year: 2022, Volume and Issue: 116, P. 105490 - 105490

Published: Oct. 7, 2022

Language: Английский

Citations

45