Theoretical and Empirical Analysis of the Learning Rate and Momentum Factor in Neural Network Modeling for Stock Prediction

Abstract

Neural Network training requires a large number of learning epochs. An appropriate learning rate is important to the overall performance of the training. Under a weight-update algorithm, a low learning rate would make the network learning slowly, and a high learning rate would make the weights and error function diverge. To optimize the model parameters, this paper presents theoretical and empirical analysis of learning rate in neural network modeling for its application in stock price prediction, an increasing learning rate approach is suggested for practice. The effect of momentum factor is also investigated to speed up the convergence for network training.