Comparative Analysis of Models for Predicting the Value of IT Companies' Shares Using Machine Learning

Stock market forecasting is a philosopher's stone for data scientists who are motivated not so much by the pursuit of material gain as by the task itself. The daily rise and fall of the market suggest that there must be patterns that we or our models can learn. There are a large number of analysis tools that allow you to make a forecast. Today, one of the popular tools for identifying short-term patterns are financial bots and machine learning models. Analysts are also interested in finding longterm dependencies in order to provide scientific justification for various economic phenomena. Neural networks are a vast family of deep learning models, and their use in finance and economics has grown significantly in recent years. The difficulty of forecasting the stock market lies in the fact that to implement high-quality forecasts, you must have complete and reliable information, as well as be able to correctly interpret it and apply it to forecasting methods. Specialized brokers base their examination on the reason that market value patterns would rehash the same thing later on, permitting them to be utilized for gauging. Time series is the most famous and very much utilized method for making expectations. The purpose of this work is to develop models of machine learning methods for modeling and forecasting the cost indicators of financial assets, conducting computer experiments and comparative analysis of the effectiveness and estimates of the forecast accuracy obtained by various models .