MULTI DATASET CLIMATOLOGICAL FORECASTING USING PSO TUNED RNN WITH MISSING VALUE IMPUTATION
DOI:
https://doi.org/10.33480/jitk.v11i4.7938Keywords:
Climatology Forecasting, Data Imputation, Missing Values, RNN-PSO, Time Series ForecastingAbstract
Missing values in climatological data due to technical glitches or extreme conditions can degrade prediction accuracy and result in biased analysis. This study aims to evaluate six imputation techniques, namely Mean, Mode, Median, Multiple Imputation by Chained Equations (MICE), Last Observation Carried Forward (LOCF), and K-Nearest Neighbors (KNN), on forecasting performance using the Recurrent Neural Network model optimized with Particle Swarm Optimization (RNN-PSO). Evaluation was carried out on three climatological time series datasets with different missing-value characteristics. The research methods include data exploration, imputation of missing values, hyperparameter optimization using PSO, and RNN model training. Model performance was evaluated using MAPE, RMSE, and R². The results show that imputation methods play an important role in maintaining the temporal integrity of the time series and consistently improving prediction precision, compared to missing data removal strategies that tend to disrupt temporal patterns and reduce model accuracy. In Dataset 1, the Median was effective in suppressing the relative error (MAPE 0.93374), but LOCF showed a more comprehensive performance by producing the lowest RMSE (3.55282) and the highest R² (0.91478), indicating its superiority in preserving temporal structure and improving overall prediction accuracy LOCF also showed strong performance on Dataset 2, especially on RMSE and R². Meanwhile, KNN provided the best results in Dataset 3 across all evaluation metrics by utilizing local correlations between data. These findings confirm that the selection of imputation methods needs to be adjusted to the characteristics of missing values, variability, and temporal properties of the dataset. Therefore, the selection of imputation methods should take into account temporal patterns and data distributions. This research provides a framework for selecting appropriate imputation methods to improve the accuracy of forecasting climatological data and support the development of a more reliable weather prediction system.
Downloads
References
[1] N. T. Luchia, E. Tasia, I. Ramadhani, A. Rahmadeyan, and R. Zahra, “Performance Comparison Between Artificial Neural Network, Recurrent Neural Network and Long Short-Term Memory for Prediction of Extreme Climate Change,” Public Research Journal of Engineering, Data Technology and Computer Science, no. Issue 2, Jan. 2024, Accessed: Dec. 06, 2025. [Online]. Available: https://journal.irpi.or.id/index.php/predatecs
[2] L. E. Alejo-Sanchez et al., “Missing data imputation of climate time series: A review,” Dec. 01, 2025, Elsevier B.V. doi: 10.1016/j.mex.2025.103455.
[3] T. M. Pham, N. Pandis, and I. R. White, “Missing data: Issues, concepts, methods,” Semin. Orthod., vol. 30, no. 1, pp. 37–44, Feb. 2024, doi: 10.1053/j.sodo.2024.01.007.
[4] L. Munkhdalai, T. Munkhdalai, V. H. Pham, M. Li, K. H. Ryu, and N. Theera-Umpon, “Recurrent Neural Network-Augmented Locally Adaptive Interpretable Regression for Multivariate Time-Series Forecasting,” IEEE Access, vol. 10, pp. 11871–11885, 2022, doi: 10.1109/ACCESS.2022.3145951.
[5] S. H. Lim, “Understanding Recurrent Neural Networks Using Nonequilibrium Response Theory,” 2021. [Online]. Available: http://jmlr.org/papers/v22/20-620.html.
[6] E. A. Nketiah, L. Chenlong, J. Yingchuan, and S. A. Aram, “Recurrent neural network modeling of multivariate time series and its application in temperature forecasting,” PLoS One, vol. 18, no. 5 May, May 2023, doi: 10.1371/journal.pone.0285713.
[7] M. K. Bhatia and V. Bhatt, “Forecasting Time Series Data using Recurrent Neural Networks: A Systematic Review,” Journal for Research in Applied Sciences and Biotechnology, vol. 3, no. 6, pp. 184–189, Dec. 2024, doi: 10.55544/jrasb.3.6.22.
[8] A. Turner, “Eighty Years of Canadian Climate Data,” Kaggle. Accessed: Oct. 09, 2025. [Online]. Available: https://www.kaggle.com/datasets/aturner374/eighty-years-of-canadian-climate-data
[9] Abhinand, “Daily Sunspot Data (1818–2019),” Kaggle. Accessed: Oct. 09, 2025. [Online]. Available: https://www.kaggle.com/datasets/abhinand05/daily-sun-spot-data-1818-to-2019
[10] N. Tricky, “Sofia Weather Records,” Kaggle. Accessed: Oct. 09, 2025. [Online]. Available: https://www.kaggle.com/datasets/nikitricky/sofia-weather-records
[11] M. Afkanpour, E. Hosseinzadeh, and H. Tabesh, “Identify the most appropriate imputation method for handling missing values in clinical structured datasets: a systematic review,” BMC Med. Res. Methodol., vol. 24, no. 1, Dec. 2024, doi: 10.1186/s12874-024-02310-6.
[12] M. W. Heymans and J. W. R. Twisk, “Handling missing data in clinical research,” J. Clin. Epidemiol., vol. 151, pp. 185–188, Nov. 2022, doi: 10.1016/j.jclinepi.2022.08.016.
[13] T. M. Pham, N. Pandis, and I. R. White, “Missing data, part 2. Missing data mechanisms: Missing completely at random, missing at random, missing not at random, and why they matter,” Jul. 01, 2022, Elsevier Inc. doi: 10.1016/j.ajodo.2022.04.001.
[14] M. Yan, L. Zhou, C. Zhao, C. Shi, and C. Ou, “Comparison of different approaches in handling missing data in longitudinal multiple-item patient-reported outcomes: a simulation study,” Health Qual. Life Outcomes, vol. 23, no. 1, Dec. 2025, doi: 10.1186/s12955-025-02364-0.
[15] R. J. Little, “Annual Review of Statistics and Its Application Missing Data Assumptions,” vol. 27, p. 23, 2026, doi: 10.1146/annurev-statistics-040720.
[16] M. W. , & T. J. W. R. Heymans, “Handling missing data in clinical research.,” J. Clin. Epidemiol., 2022, doi: https://doi.org/10.1016/j.jclinepi.2022.04.017.
[17] B. Gomer and K. H. Yuan, “A Realistic Evaluation of Methods for Handling Missing Data When There is a Mixture of MCAR, MAR, and MNAR Mechanisms in the Same Dataset,” Multivariate Behav. Res., vol. 58, no. 5, pp. 988–1013, 2023, doi: 10.1080/00273171.2022.2158776.
[18] Y. Zhang and P. J. Thorburn, “Handling missing data in near real-time environmental monitoring: A system and a review of selected methods,” Future Generation Computer Systems, vol. 128, pp. 63–72, Mar. 2022, doi: 10.1016/j.future.2021.09.033.
[19] A. Gabrio, C. Plumpton, S. Banerjee, and B. Leurent, “Linear mixed models to handle missing at random data in trial-based economic evaluations,” Health Economics (United Kingdom), vol. 31, no. 6, pp. 1276–1287, Jun. 2022, doi: 10.1002/hec.4510.
[20] P. Buczak, J. J. Chen, and M. Pauly, “Analyzing the Effect of Imputation on Classification Performance under MCAR and MAR Missing Mechanisms,” Entropy, vol. 25, no. 3, Mar. 2023, doi: 10.3390/e25030521.
[21] J. H. Li et al., “Comparison of the effects of imputation methods for missing data in predictive modelling of cohort study datasets,” BMC Med. Res. Methodol., vol. 24, no. 1, Dec. 2024, doi: 10.1186/s12874-024-02173-x.
[22] K. Seu, M.-S. Kang, and H. Lee, “INTERNATIONAL JOURNAL ON INFORMATICS VISUALIZATION journal homepage : www.joiv.org/index.php/joiv INTERNATIONAL JOURNAL ON INFORMATICS VISUALIZATION An Intelligent Missing Data Imputation Techniques: A Review,” May 2022. [Online]. Available: www.joiv.org/index.php/joiv
[23] P. Buczak, J. J. Chen, and M. Pauly, “Analyzing the Effect of Imputation on Classification Performance under MCAR and MAR Missing Mechanisms,” Entropy, vol. 25, no. 3, Mar. 2023, doi: 10.3390/e25030521.
[24] M. Alwateer, E.-S. Atlam, M. M. A. El-Raouf, O. A. Ghoneim, and I. Gad, “Missing Data Imputation: A Comprehensive Review,” Journal of Computer and Communications, vol. 12, no. 11, pp. 53–75, 2024, doi: 10.4236/jcc.2024.1211004.
[25] J. H. Li et al., “Comparison of the effects of imputation methods for missing data in predictive modelling of cohort study datasets,” BMC Med. Res. Methodol., vol. 24, no. 1, Dec. 2024, doi: 10.1186/s12874-024-02173-x.
[26] L. O. Joel, W. Doorsamy, and B. S. Paul, “A comparative study of imputation techniques for missing values in healthcare diagnostic datasets,” Int. J. Data Sci. Anal., vol. 20, no. 7, pp. 6357–6373, Nov. 2025, doi: 10.1007/s41060-025-00825-9.
[27] Y. Wang et al., “Research on Missing Value Imputation to Improve the Validity of Air Quality Data Evaluation on the Qinghai-Tibetan Plateau,” Atmosphere (Basel)., vol. 14, no. 12, Dec. 2023, doi: 10.3390/atmos14121821.
[28] S. M. Memon, R. Wamala, and I. H. Kabano, “A comparison of imputation methods for categorical data,” Inform. Med. Unlocked, vol. 42, Jan. 2023, doi: 10.1016/j.imu.2023.101382.
[29] F. A. Tyas, M. Setianama, R. F. Fajriyah, and A. Ilham, “Implementation of Particle Swarm Optimization (PSO) to Improve Neural Network Performance in Univariate Time Series Prediction,” Kinetik: Game Technology, Information System, Computer Network, Computing, Electronics, and Control, no. No. 4, Nov. 2021.
[30] M. L. Lin, C. W. Tsai, and C. K. Chen, “Daily maximum temperature forecasting in changing climate using a hybrid of Multi-dimensional Complementary Ensemble Empirical Mode Decomposition and Radial Basis Function Neural Network,” J. Hydrol. Reg. Stud., vol. 38, Dec. 2021, doi: 10.1016/j.ejrh.2021.100923.
[31] M. M. Hasan, M. J. Hasan, and P. B. Rahman, “Comparison of RNN-LSTM, TFDF and stacking model approach for weather forecasting in Bangladesh using historical data from 1963 to 2022,” PLoS One, vol. 19, no. 9 September, Sep. 2024, doi: 10.1371/journal.pone.0310446.
[32] Peshawa J. Muhammad Ali, “Investigating the Impact of Min-Max Data Normalization on the Regression Performance of K-Nearest Neighbor with Different Similarity Measurements,” 2021, Accessed: Dec. 06, 2025. [Online]. Available: (doi:10.14500/aro.10955)
[33] F. Masri, D. Saepudin, and D. Adytia, “Forecasting of Sea Level Time Series using Deep Learning RNN, LSTM, and BiLSTM, Case Study in Jakarta Bay, Indonesia.”
[34] H. Song and H. Choi, “Forecasting Stock Market Indices Using the Recurrent Neural Network Based Hybrid Models: CNN-LSTM, GRU-CNN, and Ensemble Models,” Applied Sciences (Switzerland), vol. 13, no. 7, Apr. 2023, doi: 10.3390/app13074644.
[35] L. Larasati, S. Saadah, and P. E. Yunanto, “Recurrent Neural Networks (RNN) and Long Short-Term Memory (LSTM) Methods to Forecast Daily Turnover at BM Motor Ngawi,” Indonesian Journal of Artificial Intelligence and Data Mining, vol. 7, no. 1, p. 141, Jan. 2024, doi: 10.24014/ijaidm.v7i1.27643.
[36] M. K. Bhatia and V. Bhatt, “Forecasting Time Series Data using Recurrent Neural Networks: A Systematic Review,” Journal for Research in Applied Sciences and Biotechnology, vol. 3, no. 6, pp. 184–189, Dec. 2024, doi: 10.55544/jrasb.3.6.22.
[37] A. M. Assaf, H. Haron, H. N. Abdull Hamed, F. A. Ghaleb, S. N. Qasem, and A. M. Albarrak, “A Review on Neural Network Based Models for Short Term Solar Irradiance Forecasting,” Jul. 01, 2023, Multidisciplinary Digital Publishing Institute (MDPI). doi: 10.3390/app13148332.
[38] S. Khan et al., “Optimizing load demand forecasting in educational buildings using quantum-inspired particle swarm optimization (QPSO) with recurrent neural networks (RNNs):a seasonal approach,” Sci. Rep., vol. 15, no. 1, Dec. 2025, doi: 10.1038/s41598-025-04301-z.
[39] D. Kilichev and W. Kim, “Hyperparameter Optimization for 1D-CNN-Based Network Intrusion Detection Using GA and PSO,” Mathematics, vol. 11, no. 17, Sep. 2023, doi: 10.3390/math11173724.
[40] H. S. Salem, M. A. Mead, and G. S. El-Taweel, “Particle Swarm Optimization-Based Hyperparameters Tuning of Machine Learning Models for Big COVID-19 Data Analysis,” Journal of Computer and Communications, vol. 12, no. 03, pp. 160–183, 2024, doi: 10.4236/jcc.2024.123010.
[41] A. Yunita et al., “Performance analysis of neural network architectures for time series forecasting: A comparative study of RNN, LSTM, GRU, and hybrid models,” Dec. 01, 2025, Elsevier B.V. doi: 10.1016/j.mex.2025.103462.
[42] W. H. Lin, P. Wang, K. M. Chao, H. C. Lin, Z. Y. Yang, and Y. H. Lai, “Wind power forecasting with deep learning networks: Time-series forecasting†,” Applied Sciences (Switzerland), vol. 11, no. 21, Nov. 2021, doi: 10.3390/app112110335.
[43] Y. Zhang and P. J. Thorburn, “Handling missing data in near real-time environmental monitoring: A system and a review of selected methods,” Future Generation Computer Systems, vol. 128, pp. 63–72, Mar. 2022, doi: 10.1016/j.future.2021.09.033.
[44] V. Kramar and V. Alchakov, “Time-Series Forecasting of Seasonal Data Using Machine Learning Methods,” Algorithms, vol. 16, no. 5, May 2023, doi: 10.3390/a16050248.
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Adhelia Wida Khaidir, Aji Prasetya Wibawa, Dhia Rafifah Thifal, Adelia Desyana Eka Putri, Adelia Khansa Ristiaputri, Agung Bella Putra Utama
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
-a.jpg)
-b.jpg)
