<![CDATA[Expectation Maximization (EM) was the best method used to estimate the parameters of phosphorus adsorption in a nonlinear model. However, it is questionable whether the optimum value obtained was exactly a global optimum value. Genetic algorithm is an alternative procedure to estimate the phosphorus adsorption’s parameters in a nonlinear model. The objective of this study was to have a better understanding in the use of genetic algorithm in maximum likelihood estimation of phosphorus adsorption nonlinear model parameters and compare it with the EM algorithm. This study used data of P adsorption isotherms of kaolinitic and smektitic soil in three locations. Phosphorus adsorption nonlinear models used are Freundlich and Langmuir. Results showed that the genetic algorithm and EM method produced different values of estimated phosphorus maximum adsorption and bonding energy parameters. AIC and SBC values of genetic algorithm is lower than EM algorithm, both on the Langmuir and Freundlich models. AIC and SBC values of Langmuir model is lower than Freundlich model both for genetic algorithms and EM algorithm. Hence, the best model for phosphorus adsorption is Langmuir nonlinear model with genetic algorithm. Keywords: nonlinear model, EM, Freundlich, Langmuir, genetic algorithm]]>
Copyrights © 2010
