Adaptive Client Selection for Federated IoT Intrusion Detection Under Non-IID Data Distributions

The rapid proliferation of Internet of Things (IoT) and Internet of Medical Things (IoMT) devices has significantly expanded the cyberattack surface of modern networks, making effective intrusion detection essential for protecting both system security and sensitive patient data. While Federated Learning (FL) enables collaborative intrusion detection without directly sharing raw client data, its performance is often degraded by non-independent and identically distributed (non-IID) data across clients. To address this challenge, this work proposes a Multi-Criteria Client Selection (MCCS) strategy that evaluates clients using a composite score combining local accuracy gain, Macro-F1 score, and data distribution divergence. Experiments conducted on the MQTT (Message Queuing Telemetry Transport) intrusion detection dataset under Dirichlet-based non-IID distributions demonstrate that the proposed approach achieves higher overall classification performance than standard FedAvg in the evaluated experimental settings. These results highlight the importance of intelligent client selection for federated intrusion detection under heterogeneous non-IID data distributions.

Authors

Institutions

Publication Details

Journal
Sensors
Published
2026-11-19
DOI
https://doi.org/10.3390/s26227323
Primary Topic
Privacy-Preserving Technologies in Data
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Adaptive Client Selection for Federated IoT Intrusion Detection Under Non-IID Data Distributions

Ahmed Mehaoua, Osman Salem, Ming Jun Zhang, Jiahui Xiang et al.
Sensors
Privacy-Preserving Technologies in Data
article

Adaptive Client Selection for Federated IoT Intrusion Detection Under Non-IID Data Distributions

Ahmed Mehaoua, Osman Salem, Ming Jun Zhang, Jiahui Xiang, Lydia Yaker
article en

Abstract

The rapid proliferation of Internet of Things (IoT) and Internet of Medical Things (IoMT) devices has significantly expanded the cyberattack surface of modern networks, making effective intrusion detection essential for protecting both system security and sensitive patient data. While Federated Learning (FL) enables collaborative intrusion detection without directly sharing raw client data, its performance is often degraded by non-independent and identically distributed (non-IID) data across clients. To address this challenge, this work proposes a Multi-Criteria Client Selection (MCCS) strategy that evaluates clients using a composite score combining local accuracy gain, Macro-F1 score, and data distribution divergence. Experiments conducted on the MQTT (Message Queuing Telemetry Transport) intrusion detection dataset under Dirichlet-based non-IID distributions demonstrate that the proposed approach achieves higher overall classification performance than standard FedAvg in the evaluated experimental settings. These results highlight the importance of intelligent client selection for federated intrusion detection under heterogeneous non-IID data distributions.

SensorsVol. 26(22)
Université Paris Cité (FR), Dongguan University of Technology (CN), Centre Borelli (FR)
Privacy-Preserving Technologies in Data
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon