25 Environmental Site Profiling: A Comparative Study Available to Purchase

Over the years, many methods have been developed to profile environmental contaminants in soil media and∕or groundwater. These methods vary in their ability to make precise and accurate predictions. This paper highlights the differences between the following nine profiling methodologies: Inverse distance to a power, Kriging, Minimum Curvature, Modified Shepard's, Nearest Neighbor, Polynomial Regression, Radial Basis Function, Local Polynomial, and Artificial Neural Networks (ANNs). Because each method uses an individualized logic, the accuracy of the methods' predicted profiles is expected to vary. To illustrate this, a hypothetical data-rich contaminated site is used for this purpose. Accordingly, a small fraction of the available data (about 1%) is presented to each method for site profiling. A comparative study of the models' site profiling outcomes∕predictions is then performed in order to assess the most accurate site profiling methodology. Overall, ANN-based profiling methodology outperformed the performance of all other profiling methodologies considered in this study.