Data-driven models for atmospheric air temperature forecasting at a continental climate region

• Applying a Bio-inspirited algorithm to select the optimal hyperparameters of SVR • Studying to what extent the size of the data used to train the performed models affects the accuracy of predictions. This task can be accomplished using different training and testing rations. Acknowledgments The authors would like to thank the anonymous reviewers for their constructive comments, which significantly improved this research. Author Contributions Conceptualization: Adil Masood. Data curation: Mohammed Majeed Hameed. Formal analysis: Faidhalrahman Khaleel, Mohammed Abdulhakim AlSaadi, Mohammed Majeed Hameed. Funding acquisition: Mohammed Abdulhakim AlSaadi. Investigation: Mohamed Khalid Alomar, Siti Fatin Mohd Razali. Methodology: Faidhalrahman Khaleel, Adil Masood, Siti Fatin Mohd Razali. Project administration: Siti Fatin Mohd Razali, Mohammed Abdulhakim AlSaadi. Resources: Siti Fatin Mohd Razali. Software: Mohammed Majeed Hameed. Supervision: Nadhir Al-Ansari. Validation: Faidhalrahman Khaleel. Visualization: Mohammed Majeed Hameed. Writing – original draft: Adil Masood, Mohammed Majeed Hameed. Writing – review & editing: Mohamed Khalid Alomar, Faidhalrahman Khaleel, Mustafa M. Aljumaily. References 1. Woolf S. H. and Aron L., ’US. Health in International Perspective: Shorter Lives, Poorer Health’, U.S. Heal. Int. Perspect. Shorter Lives, Poorer Heal., pp. 1–394, 2013, https://doi.org/10.17226/13497 PMID: 24006554 Table 10. Comparison of the performance of the models when two different data division methods are applied. without randomization Randomization method

Model RMSE R U MAE RMSE R U MAE QRT-M9 3.691 0.961 0.13 2.835 3.659 0.962 0.130 2.828 GBR-M6 3.612 0.963 0.128 2.791 3.598 0.965 0.129 2.795 RT-M5 3.67 0.962 0.13 2.831 3.655 0.962 0.131 2.849 SVR-M5 3.592 0.964 0.127 2.745 3.582 0.963 0.129 2.760 RF-M1 1.776 0.991 0.063 1.353 1.697 0.992 0.061 1.325 https://doi.org/10.1371/journal.pone.0277079.t010 PLOS ONE Air temperature forecasting: Mutli-time scale PLOS ONE | https://doi.org/10.1371/journal.pone.0277079 November 3, 2022 28 / 31