Evaluating the Performance of Several Data Mining Methods for Predicting Irrigation Water Requirement

Recent drought and population growth are plantingunprecedented demand for the use of available limitedwater resources. Irrigated agriculture is one of the majorconsumers of fresh water. Huge amount of water inirrigated agriculture is wasted due to poor watermanagement practices. To improve water management inirrigated areas, models for estimation of future waterrequirements are needed. Developing a model forIrrigation water demand forecasting based on historicaldata is critical to effectively improve the watermanagement practices and maximise water productivity.Data mining can be used effectively to build such models.Data mining is capable of extracting and interpreting thehidden patterns from a large amount of hydrological data.In recent years, use of data mining has become morecommon in hydrological modelling.In this paper, we compare the effectiveness of sixdifferent data mining methods namely decision tree (DT),artificial neural networks (ANNs), systematicallydeveloped forest (SysFor) for multiple trees, supportvector machine (SVM), logistic regression and thetraditional Evapotranspiration (ETc) methods andevaluate the performance of these models to predictirrigation water demand using pre-processed dataset. Thepre-processed dataset we use in this study and SysForwere never used before to compare with any otherclassification techniques. Our experimental resultindicates SysFor produces the best prediction with 97.5%accuracy followed by decision tree with 96% and ANNwith 95% respectively by closely matching thepredictions for water demand with actual water usage.Therefore, we recommend using SysFor and DT modelsfor irrigation water demand forecasting..Keywords: Irrigation water demand forecasting, Datamining, Decision tree, ANN, Multiple trees and Watermanagement.