Table 1.
Overview of Federated Learning (FL) and data heterogeneity.
Fig 1.
Conventional model training.
Fig 2.
Collaborative model training (Federated learning).
Fig 3.
Methodology—Federated learning with different clients.
Fig 4.
Flowchart.
Fig 5.
Proposed study—Experimental investigation process.
Table 2.
Datasets detail.
Fig 6.
Accuracy of IID vs. non-IID (Model: CNN, Clients: 10, Rounds: 50).
Table 3.
Global accuracy using CNN.
Table 4.
Global accuracy using MLP.
Fig 7.
Accuracy of IID vs. non-IID (Model: MLP, Clients: 10, Rounds: 50).
Table 5.
Global accuracy with more clients.
Fig 8.
Accuracy of IID vs. non-IID (Model: MLP, Clients 20, Rounds: 50).
Fig 9.
Accuracy of IID vs. non-IID (Model: MLP, Clients 20, Rounds: 100).
Table 6.
Global accuracy with more communiation rounds.
Fig 10.
Accuracy of IID vs. non-IID using Dataset-II (Model: MLP, Clients: 10, Rounds: 50).
Table 7.
Global accuracy using dataset-II.
Table 8.
Global accuracy using Dataset-III.
Fig 11.
Accuracy of IID vs. non-IID using Dataset-III (Model: MLP, Clients: 10, Rounds: 20).
Fig 12.
Performance decrease comparison with respect to dataset size.
Table 9.
Comparative analysis.