Analysis of Algal Growth Using Kohonen Self Organizing Feature Map (SOM) and its Prediction Using Rule Based Expert System

Phytoplankton becomes a concern to the environment when it forms dense growth at the water surface, known as algal bloom. However, studies on mechanism of algal bloom are not straight forward mainly caused by uncertainty and complexity of alga ecosystems. This paper describes the analysis of limnological time-series of Putrajaya Lake and wetlands to determine the growth of alga based on Kohonen self organizing feature maps (SOM). It specifically concentrates on the total Bacillariophyta species due to formation of largest algal composition in the Lake Putrajaya. An expert system was then developed based on the rules extracted from the SOM to model and predict the algal growth. The effectiveness of this system was tested on an actual tropical lake data which yields an acceptable high level of accuracy.