XAI empowered dual band Wi-Fi based indoor localization via ensemble learning - Archive ouverte HAL Accéder directement au contenu
Communication Dans Un Congrès Année : 2023

XAI empowered dual band Wi-Fi based indoor localization via ensemble learning

Résumé

Wi-Fi technology is widely used in indoor positioning systems due to its ubiquitous presence in almost every building and its cost-effectiveness without requiring additional hardware. To mitigate the effects experienced by wireless networks, dual-band Wi-Fi studies have gained importance. In this study, the UTMInDualSymFi dataset is utilized to evaluate the performance of single-band and dual-band Wi-Fi localization using 2.4 GHz and 5 GHz Wi-Fi data. For localization, KNN (K-Nearest Neighbor), XGBoost, Decision Tree, and Random Forest techniques are used for classification, and a multi-view ensemble learning approach is proposed for increasing accuracy. The results are evaluated using explainable neural network models: SHAP (SHapley Additive exPlanations) and LIME (Local Interpretable Model-Agnostic Explanations), and the effectiveness of single-band versus dual-band localization is assessed, along with the contribution of each access point to localization accuracy.
Fichier principal
Vignette du fichier
NoF2023_Final.pdf (541.34 Ko) Télécharger le fichier
Origine : Fichiers produits par l'(les) auteur(s)

Dates et versions

hal-04302091 , version 1 (23-11-2023)

Identifiants

Citer

Arzu Gorgulu Kakisim, Zeynep Turgut, Tulin Atmaca. XAI empowered dual band Wi-Fi based indoor localization via ensemble learning. 14th International Conference on Network of the Future (NoF), IEEE, Oct 2023, Izmir, Turkey. pp.150-158, ⟨10.1109/NoF58724.2023.10302788⟩. ⟨hal-04302091⟩
6 Consultations
15 Téléchargements

Altmetric

Partager

Gmail Facebook X LinkedIn More