Prediction of Sea Level Oscillations: Comparison of Regression-based Approach

In Malaysia, especially in the east coast region on the peninsula, rely heavily of the sea level reading to alert for protecting the low-lying residential regions along the coastal areas. Because recent climate change has driven the rise of sea level globally, it is imperative that the government ha...

Full description

Saved in:
Bibliographic Details
Published inEngineering letters Vol. 29; no. 3; p. 990
Main Authors Jamali, Ahmad Fitri, Mustapha, Aida, Mostafa, Salama A
Format Journal Article
LanguageEnglish
Published Hong Kong International Association of Engineers 28.08.2021
Subjects
Algorithms
Coastal zone
Coasts
Data analysis
Decision trees
Lead time
Natural disasters
Regression analysis
Regression models
Root-mean-square errors
Sea level
Malaysia
Online AccessGet full text

Cover

More Information
Summary:In Malaysia, especially in the east coast region on the peninsula, rely heavily of the sea level reading to alert for protecting the low-lying residential regions along the coastal areas. Because recent climate change has driven the rise of sea level globally, it is imperative that the government has the capacity to estimate any increase in sea level with sufficient lead time in the case of natural disaster. This study primarily aims to investigate the validity and effectiveness of four regression models, which are the Decision Tree Regression (DTR), Decision Forest Regression (DFR), Linear Regression (LR), and Bayesian Linear Regression (BLR) for predicting the monthly variation of the mean sea level. Variations of the regression models are used because these techniques have not been explored to predict mean sea level in coastal areas. The input dataset is sourced from Kerteh, Tioman Island, and Tanjung Sedili in Malaysia from January 2007 to December 2017. The performance of all algorithms are measured and compared based on the class Mean Absolute Error (MAE), Root Mean Square Errors (RMSE), Relative Absolute Error (RAE), Relative Squared Error (RSE), and Coefficient of Determination (R2). The results are hoped to model predictions of the mean sea level as part of activity in performing sea level data analysis. Therefore, this research will be able to alert other government and authorities to make an early strategy to handle the problems.
ISSN:1816-093X
1816-0948