An Application of Neural Networks: Predicting Corn Yields

Neural networks consist of highly parallel, interconnected, simple processing units. These systems differ radically from conventional computing systems; no programming is required, and knowledge is stored in the topology of the net and in the connection matrix, rather than explicitly coded in defined data structures. They offer an alternative to rule-based expert systems for developing intelligent applications. Computer algorithms allow these systems to learn from examples and generalize this learned knowledge for '''each unique situation. They provide an extremely powerful method for storing and recovering relational information in symbolic and numeric domains. Neural network software was used on weather data, Modeling corn yields allows an alternative to yield projections made by other methods, and can provide an early forecast of corn yields.