Prediction of the Quality of Fresh Water in a Basin

Abstract

Derivatives play an important role in social and economic studies. They describe the behavior of conditional expectations. Once a phenomena is characterized by parametric specifications, the conditional expectation m(x) may be modeled by a regression function. Then, derivatives may be computed by fitting the regression function. In applications, parametric estimators are commonly used, because of the un-knowledge of other more effective methods. The validity of a regression fitting approach depends on the knowledge of certain aspects related with the true functional form. In this paper, we develop a study on the usage of soft computing methods for providing an alternative to the use of non-parametric regression. We develop our modeling including neural networks and rough sets approaches. The studied problem is the eutrophication due to the growth of the population of algae. Real life data is provided by a study on a fresh water basin. They are used for developing a comparison of different approaches. A methodology is recommended for implementing a monitoring system of the water quality.