A Comparison Of Artificial Neural Networks And Statistical Regression With Biological Resources Applications

Artificial neural networks (ANNs) have been increasingly used as a model for streamflow forecasting, time series prediction, and other applications. The high interest in ANNs comes from their ability to approximate complex nonlinear functions. However, the “black-box” nature of ANN models makes it difficult for researchers to design network structure or to physically interpret the variables involved. Recent investigations in ANN research have found connections linking ANNs and statistics-based regression modeling. By comparing the two modeling structures, new insight can be gained on the functionality of ANNs. This study investigates two primary relationships between ANN and statistical models: the potential equivalence between feed-forward neural networks…

Author: Resop, Jonathan Patrick

Source: University of Maryland

Download Link: Click Here To Download This Report (PDF)

Reference URL 1: Visit Now

Reference URL 2: Visit Now

Leave a Comment