CURVE FITTING FOR COARSE DATA USING ARTIFICIAL NEURAL NETWORK

摘要

This paper demonstrates the capability of curve fitting using Artificial Neural Network (ANN), not only for a moderate set of input data but also for a coarse set of input. When appropriate number of neurons is chosen for the training purpose, accurate graphs can be obtained, despite having a coarse data. The effect of number of neurons used for curve fitting and the accuracy obtained is also studied. This aspect of ANN has been illustrated through 2 examples, Weibull distribution and another complex sinusoidal system. This curve fitting technique has been applied to a real world problem i.e. mechanism of a deep drawing press, for both slider displacement and slider velocity.