EH-ZigBeeNet Dataset for Machine and Deep Learning

The EH-ZigBeeNet Dataset for Machine and Deep Learning is a curated and preprocessed version of the original EH-ZigBeeNet IoT dataset, which is publicly available on Zenodo (record ID: 18094281) [available at https://zenodo.org/records/18094281]. This dataset has been specifically prepared to facilitate the development, training, and evaluation of machine learning and deep learning models for intelligent routing at the network layer in energy-harvesting IoT environments. The original dataset was collected from an energy-harvesting ZigBee-based mesh IoT network, where sensor nodes operate under dynamic energy constraints and communicate with a central coordinator. To improve usability for data-driven modeling, all individual data files from the original repository have been merged into a single consolidated dataset, ensuring consistency and ease of access for learning-based approaches. The final dataset consists of 51,336 rows and 35 feature columns, capturing network-level, link-quality, and energy-related parameters relevant to routing decision-making. These features enable the modeling of node behavior, network dynamics, and energy-aware communication patterns under realistic operating conditions. By providing a machine-learning-ready representation of EH-ZigBeeNet, this dataset supports reproducible research and accelerates the development of intelligent, energy-efficient routing protocols for next-generation IoT networks.