Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{202843,
        author = {Aditya Amar Khatavkar and Nikita Shetty and Shriman Dasharath Ghugare and Devraj Goraksh Shirsath and Nrupesh Vilas Kamble and Shubham Shivaji Shende},
        title = {Temperature Forecasting Using Federated Learning on Edge Devices},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {12},
        pages = {9195-9202},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=202843},
        abstract = {It is challenging to build a system that can provide accurate predictions in terms of temperatures using conventional machine learning techniques since they rely on huge computational capabilities. This paper presents an innovative FL-based framework for temperature prediction directly from the edge devices in Android smartphones. Contrary to the use of resource-intensive recurrent neural networks, the suggested method utilizes a small Multi-Layer Perceptron (MLP) in combination with sliding window temporal feature extraction to identify short-range weather patterns. Each node collects past weather data from the Open-Meteo API, pre-processes and normalizes the data locally, and trains the MLP locally as well. Only the trained weights of the model are transferred to a Python-based server that aggregates the models using the Federated Averaging (FedAvg) technique to generate the global model. The system was evaluated on three mobile phones placed in three different locations in the state of Maharashtra, namely Karad, Satara, and Umbraj. The simulation outcomes indicate steady learning behavior, obtaining a MAE value of 0.245°C and MSE value of 0.1035 for a 24-hour forecast horizon.},
        keywords = {Edge computing, federated averaging, federated learning, multi-layer perceptron, temperature forecasting.},
        month = {May},
        }