Optimisation Algorithms for Deep Learning Method: A Review with a Focus on Financial Applications
Synopsis
In a variety of fields including financial applications like stock market analysis deep learning has achieved amazing success in producing precise forecasts. To train deep learning models for financial forecasts, however, is a difficult undertaking that calls for careful consideration of a variety of hyperparameters and optimisation strategies. Optimisation is a technique that is part of mathematics and is used to solve analytical and numerical problems in minimisation and maximisation of functions. It is thus used for getting improved prediction in terms of quality and performance. In this paper we discuss different techniques like SGD, AdaGram and others, that have proven effective in improving the convergence and generalization performance of deep learning models in finance. Here we focus on financial applications where deep learning algorithms are used for the problem solving were optimization is also a part.


This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.