Application of Back-Propagation Neural Network in Horoscope Prediction

Abstract

In this study a back-propagation neural network model is designed and its parameters are optimized for prediction of horoscope to identify a person type. Person type is a dynamic system based on the planet system. It is found that the backpropagation neural network is capable to predict the person type by learning planet dataset. The model is trained up to model error (i.e., mean square error) 1.2864E-04 and performs excellent during training and testing process.