Prediction of Population Dynamics of Bacillariophyta in the Tropical Putrajaya Lake and Wetlands (Malaysia) by a Recurrent Artificial Neural Networks

Phytoplankton becomes a concern to the society when it forms a dense growth at water surface known as algae bloom. This paper discusses feasibility of applying recurrent artificial neural network to predict occurrence of selected phytoplankton population the Bacillariophyta population in Putrajaya Lake and Wetlands for one month ahead prediction. The data used are monthly data collected from August 2001 until May 2006. Network performance is measured based on the root mean square error value (RMSE). Input selection is carried out by means of correlation analysis, sensitivity analysis and unsupervised neural network SOM. Better results are achieved for simpler network where variables are selected using method stated above. Thus the capability of neural network model as a predictive tool for tropical lake cannot be disregarded at all.