Machine learning models for predicting a daily maximum water level in a tidal river: A Kapuas Kecil River case study

Sampurno, Joko;Hanert, Emmanuel;et.al.
(2021) the International Conference on Radioscience, Equatorial Atmospheric Science and Environment (INCREASE) — Location: Indonesia (20.September.2021)

Files

INCREASE2021_JokoSampurnoetal.pdf
  • Open Access
  • Adobe PDF
  • 1.63 MB
Download

Details

Authors
Abstract
In a tidal river, water level modelling is essential to assess the hazards of compound inundation. The modelling could be applied to help coastal communities to understand their risk to inundation hazards and to take adequate measures. Machine learning is a technique that is reliable for water level modelling because it can capture and represent complex input and output relationships of natural phenomena using only historical data. Therefore, this study aims to evaluate machine learning algorithms that are capable of predicting a daily maximum water level in a tidal river, particularly in Kapuas Kecil River, which runs through Pontianak. Then we select the most suitable model to assess the future flood hazards along the riverbanks. Different machine learning models, i.e., Multi-Linear Regression, Random Forest, Neural Network, and Support Vector Machine, have been evaluated. The result shows a ranking of features with the most decisive influence on the river's daily maximum water level. Then, based on the performance to predict the inundation, we propose that the support vector machine algorithm with a polynomial kernel function is the most suitable model for predicting future flood hazards in the study area.
Affiliations

Citations

  • APA
  • Chicago
  • FWB

Sampurno, J., Hanert, E., & et al. (2021). Machine learning models for predicting a daily maximum water level in a tidal river: A Kapuas Kecil River case study. the International Conference on Radioscience, Equatorial Atmospheric Science and Environment (INCREASE), Indonesia. https://hdl.handle.net/2078.5/240793