TIME SERIES ANALYSIS ON MALAYSIAN SEA LEVEL BY USING CHAOTIC APPROACH: A CASE STUDY OF SABAH

Abstract

Predicting the sea water level is very important as an increasing of sea level may lead to many disaster such as flood, inundation and coastal erosion. This research shows the application of chaotic approach for Malaysian Borneo (Sabah) sea level. The studied data, is a time series of observed sea level by hourly at the selected benchmark station located in the district of Sandakan, Kudat and Kota Kinabalu, Sabah respectively. This study has two objectives, which is to detect the presence of chaotic dynamics in data and to predict the data. To achieve the first objective, multidimensional phase space is reconstructed using the parameters of the delay time and , in Kota Kinabalu with embedding dimension d = 6 and d = 3. For Kudat, and with embedding dimension d = 5 and d = 3, and for Sandakan and with embedding dimension d = 5 and d = 4, were used respectively. Both time delay and embedding dimension values was derived from the Cao method and Average Mutual Information. The results from the phase space plot and Cao method will shows that the data are chaotic or not. To achieve the second objectives, 1 hour ahead forecasting for an hourly sea level time series is carried out by using the Local Linear Approximation method and correlation coefficient is used for testing the prediction performance. However, this method required to use MATLAB for forecast the data. The result shows the correlation coefficient value between the actual and foretasted data is 0.9 which is near to 1. This shows the reliability of the local approximation method to forecast the time series of sea level rise and also shows a positive sign that this chaotic approach is applicable to the time series of sea level in Malaysia.