Table 1 Comparison of the proposed approach with state-of-the-art diagnostic algorithms.

From: Advanced federated ensemble internet of learning approach for cloud based medical healthcare monitoring system

Criteria

FDEIoL

Algorithm A (e.g., Deep Learning Model)

Algorithm B (e.g., Centralized Model)

Algorithm C (e.g., Hybrid Model)

Efficiency

High: Utilizes edge computing for real-time processing and federated learning for scalability.

Medium: Requires high computational resources for large datasets.

Low: Centralized processing with significant latency.

Medium-High: Hybrid approach with some decentralization.

Accuracy (CXR Data)

99.24%

98.5%

97.2%

98.8%

Accuracy (Brain Tumor Data)

99.0%

98.2%

96.7%

98.4%

Resistance to Attacks

High: Incorporates federated learning, secure aggregation, and anomaly detection.

Medium: Vulnerable to adversarial attacks, limited defenses.

Low: High vulnerability due to centralized data storage.

Medium-High: Some resistance due to hybrid model structure.

Scalability

High: Scales efficiently across multiple IoT devices and data sources.

Medium: Limited by centralized processing power.

Low: Scalability issues due to centralization and bandwidth limitations.

Medium: Partially scalable but constrained by centralized components.

Data Privacy

High: Data remains decentralized with federated learning, enhancing privacy.

Medium: Requires significant data aggregation, potential privacy concerns.

Low: Centralized data storage poses high privacy risks.

Medium-High: Hybrid structure offers some privacy protection.

Data Privacy

High: Data remains decentralized with federated learning, enhancing privacy.

Medium: Requires significant data aggregation, potential privacy concerns.

Low: Centralized data storage poses high privacy risks.

Medium-High: Hybrid structure offers some privacy protection.

Model Update Frequency

High: Continuous updates from local models ensure an up-to-date global model.

Medium: Requires periodic retraining for updates.

Low: Infrequent updates due to centralized retraining needs.

Medium: Hybrid model allows for moderate update frequency.

Back to article page