Application of self-learning machines in predicting Great Salt Lake water level

Self Learning Machines (SLMs) are relatively new statistically-based methods that are becoming widely used in water resources and hydrology applications. SLMs can be trained to predict the behavior of real-time phenomena without the need of a physical model to relate the causes and effects of the phenomena. This makes SLMs an economic technique for water resources analysis and management. In this research, Great Salt Lake water level was predicted using Artificial Neural Networks (ANN), Support Vector Machines (SVMs), and Relevance Vector Machines (RVMs) as tools of SLMs. The performance of the SLMs in predicting water levels was evaluated afterwards by comparing their output to actual water levels by means of a simple statistical analysis. SLMs were applied to the Great Salt Lake, Utah, biweekly water levels as a case study. The main goal of this work was to evaluate the ability of the SLMs to predict the water level of the Great Salt Lake, in order to help managers make the right decision in case of a sudden change in the water level.