Applications of Artificial Neutral Networks in Mushroom Edibility Classification

Abstract

We report the accuracy of a two-layer, back-propagation artificial neural network in identifying edibility of a set of random mushrooms. Mushrooms edibility was synthesized using many different characteristics. Tests were run using different combinations of number of hidden nodes, separation of training, validation, and test data and number of iterations. Qualitative identification of an optimal combination of network parameters will provide a basis toward applications of artificial neural networks in future civil engineering endeavors.

Keywords

• artificial neural networks
• synthesis
• neurology