| Abstrakt: |
In recent times, Artificial Intelligence (AI) and Internet of Things (IoT) approaches are becoming the key driver for many healthcare applications. Healthcare applications are considered a significant tool that can generate a huge dataset for human diagnosis in real-time scenarios for critical cases with the support of powerful Artificial Intelligence (AI). Moreover, the effectiveness of AI model is the backbone of human diagnosis under critical conditions. Therefore, this study presents an intelligent and efficient AI-based human diagnosis framework using new hybrid optimization and machine learning models for human datasets under critical symptoms such as COVID-19 and heart diseases. The proposed hybrid framework utilizes a combination of optimization algorithms such as Firefly Algorithm (FA), Genetic Algorithm (GA), Particle Swarm Algorithm (PSO), Bat Algorithm (BA), Flower Pollination Algorithm (FPA), Jaya Algorithm (JA), and Cuckoo Search (CS), and Machine Learning (ML) such as LR, LDA, NB, KNN, and SVM. The proposed hybrid framework is verified according to two different human datasets which are Coswara-Dataset which depends on the cough sounds for COVID-19 diagnosis and The Cleveland HD dataset for heart diseases. From the experimental results, the hybrid FA-ML within the proposed framework gives acceptable results in selecting the most relevant subset of features from the two employed datasets for human diagnosis by raising the accuracy rates of the original set of features with 99.4% and 91.7% for COVID-19 and heart diseases datasets, respectively. Consequently, the results achieve superior and accurate performance compared to the other existing models for disease diagnosis in real-time based on an IoT-cloud framework. Data will be collected and analyzed via a smartphone application for more scalability and immediate assessment. [ABSTRACT FROM AUTHOR] |
Full text from SpringerLink