Diagnostic evaluation of high-impact features on intrusion detection performance in wireless sensor networks

Intrusion detection in WSNs is important to ensure network safety by identifying and counteracting unauthorized behavior. Given the limited resource nature of WSNs, finite energy, limited bandwidth, and processing power, effective and reliable intrusion detection system design is challenging. Machine learning (ML) has recently gained significant attention as an effective means to improve detection rates by recognizing complex and multi-linear patterns within network traffic. This paper investigates the use of ML regression models, i.e., Radial Basis Function (RBF) and Light Gradient Boosting Machine (LGBM), to detect intrusion events in WSNs. To increase prediction accuracy and reduce error rates, the Rider Optimization Algorithm (ROA) was used as an optimization tool to create optimized versions of these models. The performance of the optimized models was tested and validated using sensitivity analysis and SHAP (SHapley Additive exPlanations) values to detect and analyze the highest impact features influencing model outcomes. An extensive analysis of network parameters was carried out to evaluate their relative impact on intrusion detection accuracy. The analysis depicted that the Sensing Range was of the highest impact, contributing most to the prediction capability of models as reflected by high SHAP values in both LGB and RBF models. Furthermore, the Area exhibited an average degree of impact, making resource allocation strategy decisions even smarter to direct resources. The combination of strong models and interpretability sensitivity analysis was of immense value in understanding feature importance and, hence, in formulating smarter and more flexible intrusion detection mechanisms suited to resource-limited WSN environments.

Authors

Institutions

Publication Details

Journal
Journal of Engineering and Applied Science
Published
2026-11-19
DOI
https://doi.org/10.1186/s44147-026-01321-y
Primary Topic
Security in Wireless Sensor Networks
Type
article
Field-Weighted Citation Impact
0.00
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

Diagnostic evaluation of high-impact features on intrusion detection performance in wireless sensor networks

Cheng Zha, Lei Yang, Chunbaixue Yang
Journal of Engineering and Applied Science
Security in Wireless Sensor Networks
article

Diagnostic evaluation of high-impact features on intrusion detection performance in wireless sensor networks

Cheng Zha, Lei Yang, Chunbaixue Yang
article en

Abstract

Intrusion detection in WSNs is important to ensure network safety by identifying and counteracting unauthorized behavior. Given the limited resource nature of WSNs, finite energy, limited bandwidth, and processing power, effective and reliable intrusion detection system design is challenging. Machine learning (ML) has recently gained significant attention as an effective means to improve detection rates by recognizing complex and multi-linear patterns within network traffic. This paper investigates the use of ML regression models, i.e., Radial Basis Function (RBF) and Light Gradient Boosting Machine (LGBM), to detect intrusion events in WSNs. To increase prediction accuracy and reduce error rates, the Rider Optimization Algorithm (ROA) was used as an optimization tool to create optimized versions of these models. The performance of the optimized models was tested and validated using sensitivity analysis and SHAP (SHapley Additive exPlanations) values to detect and analyze the highest impact features influencing model outcomes. An extensive analysis of network parameters was carried out to evaluate their relative impact on intrusion detection accuracy. The analysis depicted that the Sensing Range was of the highest impact, contributing most to the prediction capability of models as reflected by high SHAP values in both LGB and RBF models. Furthermore, the Area exhibited an average degree of impact, making resource allocation strategy decisions even smarter to direct resources. The combination of strong models and interpretability sensitivity analysis was of immense value in understanding feature importance and, hence, in formulating smarter and more flexible intrusion detection mechanisms suited to resource-limited WSN environments.

Journal of Engineering and Applied ScienceVol. 73(1)
Pennsylvania State University (US), Anhui University of Finance and Economics (CN)
Security in Wireless Sensor Networks
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