Interpretable Investigation of Feature Relevance and Sparsity of IoT Datasets
Saroj Bala, Kumud Arora · 2025
The rapid growth of the Internet of Things has led to the generation of vast amounts of data, which presents both challenges and opportunities for researchers and practitioners. Machine learning and deep learning techniques have emerged as powerful tools for extracting valuable insights from IoT datasets. This paper presents an empirical investigation of various ML and DL models to uncover the feature relevance in IoT datasets. The review highlights the strengths and limitations of these techniques and discusses the challenges associated with their application in the IoT domain. It explores LASSONET for feature sparsity and explainable AI techniques such as LIME, SHAP and ELlS, for feature relevance. The findings of this study can inform the development of more effective and interpretable IoT analytics solutions, ultimately enabling better decision-making and optimization of IoT systems.