Please use this identifier to cite or link to this item:
http://hdl.handle.net/20.500.11960/4016
Title: | Forecasting short-term indoor radon: a machine learning approach using LSTM networks |
Authors: | Mpinga, Valdo Cruz, António Miguel Lopes, Sérgio Ivan |
Keywords: | LSTM Bi-LSTM Forecasting IoT Radon |
Issue Date: | 2023 |
Citation: | Mpinga, V., Cruz, A. M. R., & Lopes, S. I. (2023). Forecasting short-term indoor radon: a machine learning approach using LSTM networks. In Proceedings of 18th Iberian Conference on Information Systems and Technologies, CISTI, 20-23 June,2023, Aveiro (Portugal). https://doi.org/10.23919/CISTI58278.2023.10211807 |
Abstract: | Indoor radon is a radioactive gas that can accumulate in homes and pose a health risk for humans. Forecasting indoor radon levels may be used as a tool for mitigating human exposure risk, and thus help to effectively manage indoor radon risk. Forecasting based on Machine Learning (ML) techniques involves predicting future levels of indoor radon gas based on past and current data, and thus help identify trends and patterns in the data over time. This work presents preliminary results regarding the implementation and evaluation of two LSTMbased approaches, for indoor radon forecasting, which can then be used as a tool to trigger preventive management procedures for Indoor Air Quality management. Preliminary results have shown that the normalized data using the Long Short-Term Memory (LSTM) algorithm proved to be the optimal approach for this application case, demonstrating superior accuracy across various forecasting time windows when compared to other approaches evaluated in this work. |
URI: | http://hdl.handle.net/20.500.11960/4016 |
ISBN: | 978-989-33-4792-8 |
ISSN: | 2166-0727 |
Appears in Collections: | ADiT-Lab - Artigos indexados à WoS/Scopus ESTG - Artigos indexados à WoS/Scopus |
Files in This Item:
File | Description | Size | Format | |
---|---|---|---|---|
Forecasting_Short-Term_Indoor_Radon-A_Machine_Learning_Approach_Using_LSTM_Networks.pdf | 566 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.