Lake Chini Water Level Prediction Model using Classification Techniques

Abstract

Monsoon seasons in Malaysia bring uneven distribution of rainfall and eventually affect the water level at Lake Chini as flood and drought disturb the population and distribution of aquatic organisms at the lake. This study is conducted to produce Lake Chini water level prediction model by comparing several algorithms using data mining approach via classification techniques. Data from seven observation stations between 2011 and 2014 are collected from Pusat Penyelidikan Tasik Chini, Universiti Kebangsaan Malaysia and data from Melai station in particular is used for this purpose. The collected time series data is complex and high in dimensionality thus leading to low efficiency in data mining process. The analysis comprises of four phases that include data collection, data pre-processing, data mining and model development and interpretation and evaluation of patterns. To overcome high dimensional time series, dimensionality reduction approach such as Piecewise Aggregate Approximation (PAA) and Symbolic Aggregate approXimation (SAX) are applied while three classification techniques namely Decision Tree, Artificial Neural Network and Support Vector Machine are used to classify the data. Performance measures for each of the algorithms are evaluated and compared to select the most suitable model for the prediction.