Analysis and predictive validity of Kelantan River flow using RQA and time series analysis

Abstract EN

An analysis of the Kelantan River flow and its predictive validity has been under taken for the period 2000 to 2014 by utilizing Recurrence Plot and Recurrence Quantification Analysis.

The study presents the results of the application of nonlinear and linear time series analysis on the river stream-flow data for flood detection and prediction of future values.

The prediction results on a hold-out sample of 2014 data by the methods based on RQA and ARIMA have been compared and it was revealed that the ARIMA model provided a better forecast.

The analysis of the dynamics of the daily river flow and the detection of an attractor with outliers indicating higher values of the series are believed to be two important factors of an onset of a sudden change.