Modelling the underground water level via artificial nerves

This study aims to determine the behaviour of the underground water via Artificial Nerves Net (ANN) which the dependent variable flow (stream), independent variable, heat, rain, evaporation and the level of the well are used.Artificial nerves net is a computer system that can automatically fulfill some human abilities such as creating information from learning, getting new information and discovering like a human brain without any help.Due to including chemical salts, underground water damages building bases and because of capillarity it damages liquids and concretre, also. Particularly in high underground water level areas, high underground water level causes consolidation on clayey areas and displacement of buildings, so it damages the buildings. Furthermore, out of building areas, it leads unnecessarry wettness and impacts public health. We will use this model to evaluate the underground water level, which the most important criterion in building field ANN and Regression models have been examined. Comparing ANN with Regression model, it is clear that ANN concludes beter results. At the end of the model study, it is seen that the ANN model, which applies Yearly Mean, Monthly Ordinal Number, Yearly Ordinal Number and Monthly Normalizing Mean as input and Monthly Mean as output, is successful.