Аналіз можливості прогнозування цін

Abstract

This article discusses the situation of deep learning models in the stock market in our time, describes the problems that arise when using these models, the reasons for their occurrence, and ways to prevent them. It is justified that mainly should be used historical price information to build a predictive model rather than other information from various sources. There have been proposed algorithms for use and implementation. Mostly on the simplest algorithms due to their speed of analysis and, as a result, efficiency. And as a result, a conclusion is drawn on development trends.