Performance Evaluation of Ensemble Decision Tree in Identification of Metal Contaminants in Water Reservoirs

Abstract

Machine Learning Methods are getting popular day by day and finds its applications in all fields. Researchers are involved in developing new machine learning algorithms by optimizing the various parameters. On the other hand, studies are also in progress in analyzing the performance of the developed machine learning algorithms in the various applications. This paper discusses one such study done by applying decision tree and ensemble decision tree methods to identify the type of metal contaminants in the water reservoirs. The colour histogram features extracted from the images of Lemna minor was grown on waters with different metal dissolvent were used for the study. The performance evaluation done on both the methods concluded that the ensemble method shows better performance than the decision tree method with overall accuracy of 94.5% in the classification of metal contaminants in water.