Comparing Machine Learning Methods for Early Warning  of Floods and Landslides in Thailand

Joanna  Sophie Abraham; Supatta  Labaiusuh; Ekkasit  Ismael; Chatree  Nilnumpetch

doi:10.14456/nujst.2023.37

Authors

Joanna Sophie Abraham Department of Computer Science, Faculty of Science, Ramkhamhaeng University, Bangkok, 10240, Thailand
Supatta Labaiusuh Department of Computer Science, Faculty of Science, Ramkhamhaeng University, Bangkok, 10240, Thailand
Ekkasit Ismael Department of Computer Science, Faculty of Science, Ramkhamhaeng University, Bangkok, 10240, Thailand
Chatree Nilnumpetch Department of Computer Science, Faculty of Science, Ramkhamhaeng University, Bangkok, 10240, Thailand

DOI:

https://doi.org/10.14456/nujst.2023.37

Keywords:

Machine Learning, Random Forest, Decision Tree, Early Warning, Landslides

Abstract

Flood disasters and landslides have a strong impact on people's lives, property, and the economy of the country. Heavy rainfall is the primary cause of these disasters. Therefore, prediction warnings is necessary for people to help them prepare for the disaster in time. This paper outlines the process used to identify appropriate models for prediction warnings for floods and landslides by comparing the recall performance of eight different models. The models were Rule-Based, K-nearest Neighbor, Decision Tree, Random Forest, Support Vector Machine, Naïve Bayes, Logistic Regression, and Multilayer Perceptron. The process involved five phases: data collection, data pre-processing, building a model, 5-fold cross-validation, and model evaluation. This study utilized a rainfall-related dataset collected by the Department of Water Resources in Thailand for training and testing the models. After the process was applied along with a detailed evaluation, it found that when 5-fold cross-validation was applied, better performance was achieved with Random Forest having the highest recall value at 74%, followed by Decision Tree, Multilayer Perceptron and Support Vector Machine. From these results, it can be concluded that the Random Forest model is suitable for predicting warnings and can be implemented in future works for developing an early warning application to reduce the aftermath of these disasters.

References

Acharya, V., Ghosh, A., Kang, I., Munasinghe, T., & Binita, K. (2022). Landslide Likelihood Prediction using Machine Learning Algorithms. 2022 IEEE International Conference on Big Data (Big Data), 17-20 December 2022 (pp. 5395-5403). Osaka, Japan: Institute of Electrical and Electronics Engineers. http:dx.doi.org/10.1109/BigData55660.2022.10020433

Adnan, M., Rahman, M., Ahmed, N., Ahmed, B., Rabbu, M., & Rahman, R. (2020). Improving Spatial Agreement in Machine Learning-Based Landslide Susceptibility Mapping. Remote Sensing, 12, 3347. http:dx.doi.org/10.3390/rs12203347

Akter, T., Sadman, Y., & Bala, S. (2022). Use Machine Learning Technologies in E-learning. 2022 IEEE International IOT, Electronics and Mechatronics Conference (IEMTRONICS), 1-4 June 2022 (pp. 1-4). Toronto, ON, Canada: Institute of Electrical and Electronics Engineers. http:dx.doi.org/10.1109/IEMTRONICS55184.2022.9795843

Chang, K.-T., Hwang, J.-T., Liu, J.-K., Wang, E.-H., & Wang, C.-I. (2011). Apply two hybrid methods on the rainfall-induced landslides interpretation. 2011 19th International Conference on Geoinformatics, 24-26 June 2011 (pp. 1-5). Shanghai, China: Institute of Electrical and Electronics Engineers. http:dx.doi.org/10.1109/GeoInformatics.2011.5980950

Cover, T., & Hart, P. (1967, January ). Nearest neighbor pattern classification. IEEE Transactions on Information Theory, 13(1), 21-27. http:dx.doi.org/10.1109/TIT.1967.1053964

Department of Disaster Prevention and Mitigation. (2022). Popho phoei sarup sathiti satharanaphai nai rop pi 64 [DDPM reveals the summary of disaster statistics in the year 2021]. Retrieved from http://relation.disaster.go.th/

Department of Water Resources, Ministry of Natural Resources and Environment. (2009). Rai-ngan kansueksa wichai patchai siang lae withichiwit chumchon nai phuenthi siangphai din thalom lae nam lak - din thalom [Research report: risk factors and community lifestyles in landslide prone areas and floods-landslides]. Bangkok: Department of Water Resources.

Habibie, M., & Nurda, N. (2022). Performance Analysis and Classification using Naive bayes and Logistic Regression on Big Data. 2022 1st International Conference on Smart Technology, Applied Informatics, and Engineering (APICS), 23-24 August 2022 (pp. 48-52). Surakarta, Indonesia: Institute of Electrical and Electronics Engineers. http:dx.doi.org/10.1109/APICS56469.2022.9918793

Inruang, W., & Chaipimonplin, T. (2558, January-December). The Prediction of Landslides Risk Areas in Uttaradit Province by Applying Geo-Informatics Technology with an Artificial Neural Network. Journal of Social Sciences Srinakharinwirot University, 18(18), 191-207.

Jinsikhong, P., Musika, K., Amonlak, W., & Nonthakarn, D. (2022). Rabop chaeng tuean phai namthuam lae prap radap nam thang lot umong [Flood warning system and water level adjustment in the tunnel]. The 10th Asia Undergraduate Conference on Computing : AUC2, 24 February 2022 (pp. 2473-2474). Chon Buri: Kasetsart University Sriracha Campus.

Kotsiantis, S. (2013). Decision trees: a recent overview. Artificial Intelligence Review, 39, 261-283. http:dx.doi.org/10.1007/s10462-011-9272-4

Martin, D., & Chai, S. (2022). A Study on Performance Comparisons between KNN, Random Forest and XGBoost in Prediction of Landslide Susceptibility in Kota Kinabalu, Malaysia. 2022 IEEE 13th Control and System Graduate Research Colloquium (ICSGRC), 23 July 2022 (pp. 159-164). Shah Alam, Malaysia: Institute of Electrical and Electronics Engineers. http:dx.doi.org/10.1109/ICSGRC55096.2022.9845146

Naviamos, M., & Niguidula, J. (2020). A Study on Determining Household Poverty Status: SVM Based Classification Model. 3rd International Conference on Software Engineering and Information Management (ICSIM '20) 12-15 January 2020 (pp. 79-84). New York, NY, USA: Association for Computing Machinery. http:dx.doi.org/10.1145/3378936.3378969

Phomsak, W., & Khruahong, S. (2021). Rabop tuean fai pa attanomat duai theknoloyi lo ra [Automatic Forest Fire Alarm System with LoRa Technology]. The 9th Asia Undergraduate Conference on Computing : AUCC, 25–26 February 2022 (pp. 1032-1038). Prachuap Khiri Khan: Rajamangala University of Technology Rattanakosin Wang Klai Kangwon Khet Campus.

Pomthong, S., & Asavasuthirakul, D. (2017). Kan wikhro phuenthi siangphai chak din thalom nai changwat phetchabun [Analysis of Landslide Risk Area in Phetchabun Province]. Journal of Geoinformation Technology of Burapha University, 2(3), 41-52.

Rangsiwanichpong, P. (2020). Application of Geographic Information System with the Probability of Landslide Model for Assessing Landslide Hazard in the Eastern Thailand. The Journal of King Mongkut's University of Technology North Bangkok, 30(4), 560-569.

Solehman, M., Azmi, F., & Setianingsih, C. (2019). Web-Based Flood Warning System Using Decision Tree Method. 2019 International Conference on Advanced Mechatronics, Intelligent Manufacture and Industrial Automation (ICAMIMIA), 9-10 October 2019 (pp. 1-6). Batu, Indonesia: Institute of Electrical and Electronics Engineers. http:dx.doi.org/10.1109/ICAMIMIA47173.2019.9223370

Sreevidya, P., Abhilash, C., Paul, J., & Rejithkumar, G. (2021). A Machine Learning-Based Early Landslide Warning System Using IoT. 2021 4th Biennial International Conference on Nascent Technologies in Engineering (ICNTE), 15-16 January 2021 (pp. 1-6). NaviMumbai, India: Institute of Electrical and Electronics Engineers. http:dx.doi.org/10.1109/ICNTE51185.2021.9487669

Thaiyuenwong, S., Nuthimthong, P., Detmak, W., & Soralump, S. (2013). Kan tuean phai din thalom chak kan truat wat pariman nam fon nai phuenthi phukhao thang phaknuea khong prathet thai [Landslide Warning by Rainfall Monitoring on Mountainous Area in Northern Thailand]. The 18th National Convention on Civil Engineering, 8-10 May 2013 (pp. 401-406). Chiangmai: Chiang Mai University.

Wongka, M., Montri, W., & Chaikhamwang, S. (2022). Khrueang tonbaep chaeng tuean fai pa doi chai theknoloyi khrueakhai rai sai lo ra [Wildlife Alarm Model Using LoRaWAN Technology]. The 10th Asia Undergraduate Conference on Computing : AUC2, 24 February 2022 (pp. 677-685). Chon Buri: Kasetsart University Sriracha Campus.

Yang, F.-J. (2018). An Implementation of Naive Bayes Classifier. 2018 International Conference on Computational Science and Computational Intelligence (CSCI), 13-15 December 2018 (pp. 301-306). Las Vegas, NV, USA: IEEE Computer Society. http:dx.doi.org/10.1109/CSCI46756.2018.00065

Yi, Y., Zhang, Z., Zhang, W., & Xu, C. (2019). Comparison of Different Machine Learning Models For Landslide Susceptibility Mapping. IGARSS 2019 - 2019 IEEE International Geoscience and Remote Sensing Symposium, 28 July 2019- 2 Augest 2019 (pp. 9318-9321). Yokohama, Japan: Institute of Electrical and Electronics Engineers. http:dx.doi.org/10.1109/IGARSS.2019.8898208

Yilmaz, E., Teke, A., & Kavzoglu, T. (2022). Performance Evaluation of Depthwise Separable CNN and Random Forest Algorithms for Landslide Susceptibility Prediction. IGARSS 2022 - 2022 IEEE International Geoscience and Remote Sensing Symposium, 17-22 July 2022 (pp. 5477-5480). Kuala Lumpur, Malaysia: Institute of Electrical and Electronics Engineers.http:dx.doi.org/10.1109/IGARSS46834.2022.9883954