Predictive modeling of complex mathematical functions using neural networks

This research examines artificial neural networks' flexibility to forecast complicated computational processes. We created false data from the continuous field of polynomial, exponential, logarithmic, and trigonometric functions. Splitting all function training and testing sets created homogenous neural network models. MSE was compared to test data to evaluate models. Neural networks predict sine-cosine, exponential of sine, and cubic transformation functions accurately. Neural networks may capture complex operational relationships, as shown by an ordered comparison results table. This study shows that neural networks can solve mathematical modelling problems in automated forecasting.